Parts and Compatibles

Making Components and Critical Parts

This is the second part of an examination of the problems posed by spare parts.   The main example used is the RG9-1485 separation pad for the HP Laserjet 5000 and LJ-5100 printer. The part is merely a simple example. Our aim in this section is to explore the problems of making and shipping parts, and how things are changing. For instance, is there a serious market for compatible parts? If so, what would be needed to make compatible parts correctly.

Complicated assemblies like power-bricks are the subject of another essay. Apparently there are compatible and counterfeit items out there - and an examination of one shows why they can be cheap - but poor value and perhaps even dangerous.

One of our distributors has the slogan Always use Genuine HP Spare Parts - Imitations have their limitations. Generally we would agree with that. If you are going to site to fix a printer the main cost is time, so you just want the parts to work properly first time. Using manufacturers original parts gives the maximum probability that things will go right.

Another of the distributors we use had a banner saying Always use Genuine HP Spare Parts. Amusingly they used the same banner on every page - Brother, Canon, OKI, Kyocera etc. Now it can happen that HP parts will fit a Brother printer because they are both using the same ‘engine’ but HP parts might then count as ‘compatibles’. We go into more depth on how engines, printers and parts distribution all fit together below.

Because the world is a complicated place, what is of value is often not obvious.For instance: it wasn't self evident that Intel was a great investment in 1967, that Microsoft or Apple would be in 1977, Sun or Cisco in 1987 , Google in 1997 or Twitter in 2007.Money is of no value except that it is a social convention expected to last. An American dollar has roughly the value the US government gives it, allowing a bit for what politics allows possible and what currency and bond holders, speculative traders and other governments will allow.There are some wise old saws;for instance that land is usually a good investment because they aren't making any more of it. Unfortunately commodities are prone to speculative bubbles, so for half the time they actually prove bad investments.People have rather simple, stone-age needs: food, sleep, physical well-being. To have sleep we must have a safe place. We are rather peculiar animals, not at all safe on our own; but rather effective at running and hunting as a tribe. Ultimately some must have sex, or we wouldn't be here and the human child takes a long time to mature and needs some protection from adults for perhaps ten years. So it seems that although we might sometimes think of ourselves as individuals (the US tradition of Thoreau), we must be family or tribal. What evidence pre-history gives us is that we lived in tribes of 25 to 100 people.Food, sleep, sex and socialisation would be good investments then, meeting real human needs.

Understanding Manufacturing

Schools have always had a wider mission than just to teach the "three R's". A lot of older teachers tell tales about how they first took urban children to the country, where some were shocked to find milk came from a cow. It's part of a wider disconnect between the grocery shop and the realities of farm life. In the 1960's things got worse, whilst people generally had an intellectual sense that milk came from a cow, bacon from a pig they were developing a disconnect; a romanticised view of how things work. The BSE and horsemeat crises suggest that people still have very little grasp of how the food-chain operates.

The problem is clearly worse with factories because how things are made is so much less visible and obvious. The farm has trucks full of hay and cattle feed - and possibly ammonium nitrate (used as fertilizer). A factory might have a truck marked Ethylene glycol which it turns into polyethylene naphthalate - which sounds a bit less threatening called polyester, the same chemical can be used to make a PET bottle.

People have always manufactured things. The oldest clearly defined tools found in the fossil record are chipped stones about 3.4 million years old, suggesting they were made by humanity's hominid ancestors rather than our own species. Children start arranging and examining things and sometime breaking them as they develop hand-eye coordination. As individuals today and as a species we grew up with the idea of making things.

Until quite recently, however, people had limits in what they could do. You can't fell many big trees with a stone hand-axe or make a bridge longer than the tallest tree you can cut.

We still have limitations today but fewer of them.

In 2014 the worlds longest bridge is the Danyang-Kunshan Grand Bridge, a viaduct on the Beijing-Shanghai High-Speed Railway, it is 164.8 kilometers long (just over 102 miles).

We can turn lead into gold if we want using nuclear transmutation, but the process is expensive and works more easily the other way. Somehow, turning gold into lead isn't appealing:

  • gold doesn't tarnish, is an excellent electrical conductor and is easily worked into very thin wires and sheets.
  • lead tarnishes and is a bit poisonous.

Manufacturing on any scale is quite a recent phenomenon; it also has a bit of an undeserved bad reputation.It is only quite recently that many people have had the opportunity to own several sets of clothes, to malnourish themselves by overeating and to make themselves unhappy with aspiration for rather banal consumer goods.

Globalisation in manufacturing is well known but the mechanisms less so. It isn't just cheap labour shifting jobs, although that certainly plays a part. But there were going to be fewer factories anyway. Factories tend to get bigger - or at least the machines in them have higher productivity. Modern manufacturing tends to be contracted out so factories are not dedicated to one brand. There are possibly more "things" than ever before but people seem to be converging on fewer but more sophisticated possessions.

Globalisation is complex - superficially it just means one organisation operates globally. Businesses might have multiple offices and factories operating in regional markets - but they might have a single factory producing for the whole world. There is usually one head office and one key boss. Companies tend to be large, hierarchical and often rather like the ancient empires of China, Egypt or Rome in their infighting, autocratic decision making and sudden changes of fortune.

-

Getting the world in any kind of perspective is difficult: it tends to get rather complicated and then people choose an organising principle based on politics, religion or both.

To know what sort of things fundamentally appeal we need to know our own nature. That is difficult, because we have only rudimentary ideas about how thought, intelligence, conciousness or cognition actually work. We think, but we aren't sure how. We are pretty sure we think differently from the animals, but we aren't sure why and to what end.

One way to guess at what will appeal to people is to look at their senses.As a species, we have a sweet tooth.Historically most manufacturers don't worry about it, they just look at what sells. Gradually this drifts as tastes change.

There is a question about why we make things. Man has sometimes been described as "the tool making animal". Most larger animals are clearly sentient, many animals seem self-aware, some seem rational. There are a few that seem to use sound as language. What seemed to set us apart was the use of tools; but more recently it turns out that great apes, crows and dolphins amongst others can all make tools so it isn't such a distinction.

Nothing animals have built compares with a particle accelerator, moonshot or even a bicycle. (Dolphins might have more sense - as in "So long, and thanks for all the fish".)

Life as a simple hunter-gatherer sometimes look appealing; hence the Garden of Eden story.

There was never a time when people did not make tools. Doing things, making things is a human characteristic; it might even be a requirement for well-being.

People are peculiar creatures, the two legged gait for a start. Apes and a few other creatures can do this but generally prefer not to. Our genetic close relatives the chimpanzees have very strong arms to propel themselves from branch to branch in the trees. Human arms are not so substantial as the legs; it seems we are built for running. The aim probably isn't running away because compared to many predators humans don't run fast, but they can run a long way so they probably ran prey to exhaustion. After six miles a horse has overheated

Most predators have fearsome teeth and claws; humans have rather soft, sensitive hands.

Human pre-history is sometimes divided into three parts for convenience: the Stone, Bronze and Iron Ages, based on the kind of artefacts they left. The main evidence we have of our origins are the museum pieces, there are legends and belief systems, customs and practices, family and tribal ties and enmities. There must have been moments of transcendental beauty and happiness or we could not be here. Perhaps the best record of what we once were is ourselves, our languages, cultures and the genetics which make it quite clear that people are all very closely related.

Stone Age articles seem to represent the basic human toolset. In the fossil record it is characterised by stone axes and butchery tools, then later by chisels, drills, arrowheads and needles. In practice, people were probably much more focused on plant materials and furs but few of even the most substantial timbers survive. The only metals which could be found were gold, silver and copper, which are not abundant. Metal was not utterly unknown but few would have any. There was a little meteoric iron - which wasn't very attractive. There is no written history and little discernible legacy of materials or legend. However we can infer what life must have been like from the stone tools, a few remaining fossils and from recent observation of life in remoter corners of Earth. Life is based on a hunter-gatherer economy. People have a nomadic lifestyle in tribes of perhaps 25 to 100. People in the tribe are usually quite closely related though with enough exchange with neighbouring tribes to avoid inbreeding.

It is a long period in history. The earliest evidence for stone tools dates back about 3.4 million years, to the hominids who were our ancestors. Stone tools seem to have been essential for the development and spread of Homo Erectus and are found in Africa between 2.6 and 1.7 million years ago, in Java by 1.8 million years ago, and Northern China by 1.6 million years ago. It isn't clear just how old modern humans are, about half a million to a million years old - and genetic evidence shows they did interbreed a bit with their evolutionary cousins the Neanderthals and Denisovians.

There was progress: tribes spread out across the world, they presumably had loosely defined territories but bad seasons or quarrels would drive them from their customary lands in search of something new. They had aspirations like we do and a few expressed it in cave art but they must have been largely sufficient to themselves. Perhaps a love token was a curiously formed stick or shell. Clothing seems to have existed for at least 50,000 years and perhaps as much as 170,000 years; it is difficult to tell because flexible, soft materials decay. Pottery can survive and shards dating back 30,000 years have been found in the Czech Republic and 20,000 years old in China. Craft skills date back a long way although in terms of the half million or more year existence of the human species they are very recent.

This period lasts until about 6,000 years ago but towards its end the idea of agriculture changes things.

Agriculture developed about 12,000 years ago (10,000 BC) somewhere in the "fertile crescent" stretching from Egypt to North India. There probably wasn't any specific moment of invention - "if I push these seeds into the soil there will be fruit next year". Agriculture seems to have been independently discovered in China, and again in South America. Crops that had been harvested from wild stands are planted in the expectation of a harvest; then the people settle to look after their crops and that puts serious restrictions on a nomadic lifestyle.

What triggered the discovery of agriculture is unknown. The human species in one form or another had existed for hundreds of thousands of years as hunter-gatherers.

The development coincides with climate change at the end of the ice age. The areas where agriculture developed were changing to have a long dry season; plants would respond by putting more effort into seed production than woody growth, so may already have been evolving towards their domestic form.

It may have been that human numbers in certain areas outgrew the food resources a hunter gatherer lifestyle could produce, but they managed to solve the problems. Whatever the driving force, eight founder crops appeared in the Levant after 9500 BC: emmer and einkorn wheat, then hulled barley, peas, lentils, bitter vetch, chick peas and flax. Fig trees seem to date from around the same time. By about 3000 BC Andean South America had independently developed potato, tomato, peppers, squash and several varieties of beans.

Agriculture is not wholly incompatible with gathering and hunting, but working from one place means the hunting grounds become more distant. People on open grassland domesticated animals and adopted a herdsman's lifestyle. Conflict is inevitable when a herd of beasts munch their way through someone's wheat harvest.

Agriculture has problems and benefits. There is the effort of sewing and reaping and of grinding and cooking corn. A lot of the time that people once spent looking for fruit and stalking prey is now potentially idle. So someone sitting around a campfire night after night has the opportunity to observe that where the fire was hottest those grey stones have developed a metallic sheen, and to collect the stuff and make a piece of jewellery or an arrowhead. We probably don't need a sedentary agricultural tribe to get metalwork but it must help.

Something limits prior discovery.

There is a copper age in Western Asia before the Bronze age.

Bronze Age: a period of a few hundred, perhaps thousand of years in some places where surviving artefacts were made of copper-tin alloys. It isn't entirely pre-history. Some writings exist but were rare, often obscure and most have been lost. Some places never had a "bronze age" because they did not have the materials. Europeans traded to get the new products.

Around the Eastern shores of the Mediterranean there is a brief copper age. Copper metal occurs naturally amongst some of its ores, which are sulphurated salts of the metal. Getting copper won't have been difficult, but it is a bit difficult to work and not very hard or suitable for making a blade. A copper sword is better than a wooden stake, but much more expensive and no match for bronze.

Trade across long distances is nothing new; the Bronze age couldn't have happened without trade because good sources of copper are not usually found near tin ores.

The Bronze age is an important first step. It is the first flowering of serious manufacturing. Bronze is a nice material, it is harder than copper but easier to melt and forge. It is particularly suited to casting. It is still used in sculpture.

Agriculture seems to have allowed Sumeria and Persia to develop other ideas. As well as metalworking it allowed much more. Around 4,000 years ago they invented the potter's wheel, central government, palaces and law codes. They also invented social stratification and slavery. Tribal life cannot permanently sustain a hierarchy; if a substantial number disagree they can take a different fork in a path and it would be difficult to argue. Larger societies can have a professional soldier class with metal tipped spears; that makes dissent difficult. We can know some of this because they developed writing and wrote some of it down.

Stone, Bronze and Iron ages look superficially like a progression, technologies building on one another. In general there is a progress, technology gets better. In detail a tribe or city appears, prospers then collapses. Mediterranean Bronze age societies collapsed over a period, sometimes appealing to one another for help. One reason is that their settlements had difficulty resisting raids by the Sea Peoples - a lose confederacy of tribes mounting raids in warships. It wasn't as simple as that Sea Peoples had superior technology, but the Bronze age cities such as Ugarit and Troy were vulnerable.

Iron weapons weren't better than bronze (steel can be). Iron weapons were cheaper allowing larger armies which could swarm and cut down a ruling-class chariot army. It seems that trade disruption cut the supply of tin so the bronze age societies first recycled what they could and then substituted iron where they had to.

Iron Age: is a period of a thousand years or so - arguably continuing until recently, but generally regarded as ending with recorded history.

Iron ore is relatively common and widely available compared to the copper and tin needed for bronze. However iron is difficult to work because it requires high temperature furnaces and techniques previous ages couldn't achieve. Only in very recent times has iron been surpassed in manufacturing by plastics.

The earliest iron production occurred in Anatolia around 1200 BC and spread into Asia.

The iron age is not entirely prehistoric, there are writings such as the Indian Vedas and the older parts of the Hebrew Bible. Sanskrit and Chinese literature were developing strongly.

In Europe's Middle ages wool from the English Cotswolds and Northumbrian Moors was exported to Flanders and the Hanseatic League and in return came wine, exotic spices and finished goods.

Venice grew as one of the Mediterranean's chief trade ports.

The earliest global organisations were perhaps the East India Companies, notably the British East India Company founded in 1602 and only wound up in 1874, but also the Dutch, Danish, Portuguese French and Swedish companies. A scottish venture, The Darien Company, failed and ruined many wealthy Scots leading to the Act of Union.

The East India and West India companies are colonial ventures using European weapons and shipbuilding to secure control over territory. It is to be remembered that conquest and slaughter were nothing new - Aztecs, African kingdoms, and Indian princes were just as warlike and brutal - but less systematic. At the other side of the world, the great Chinese Treasure Voyages attempted something very similar, and it seems to be their withdrawal that left a power vacuum and let the British and their imitators in.

Popular music tends toward one producer:- Elvis, Beatles, Madonna, Prince - working at a global scale.

Globalisation means half a dozen manufacturing plants - sometimes more, often fewer - producing all the goods for a brand. The big movements in production that implies started in the 1950s and accelerated in the 1980's as manufacturers found big, well educated, compliant workforces in Asia and shifted production from Europe and the US.

The trend to branded goods is less perceived.

Modern factories are usually big places; it looks as though almost anything might happen in there before the products roll out of the door in container-loads.

Factories were once "vertically integrated" and almost everything did happen inside the gates. The first big factory built by Matthew Boulton in Birmingham, England made buttons, buckles, silverware and ormolu as well as housing a mint for coins. It brought lots of skilled workers together where previously the products had been made in separate little workshops. It was also the base for Boulton and Watt, a partnership that made the first efficient steam engines and so began the industrial revolution.

Later factories of this kind included Armstrong's works in Tyneside and Carnegie's works in Pittsburgh.

Ford's River Rouge Complex, a mile and a half long and a mile wide, employed over 100,000 workers to turn raw materials into running vehicles in the 1930s. The Rouge Complex was begun in 1917; it has its own dock, ore processing, coke ovens, power plant - it had 93 buildings in all. However the site now operates more like an industrial park. There are six Ford factories. The docks and furnaces were sold, the steel plant is run by Severstal North America.

One of the largest and most complex examples of integrated production is BASF's Ludwigshafen site employing 33,000 people.

Automotive products, cars, trucks and buses, are the world's largest technology driven sector. Construction may be larger but is dominated by site based wet processes like concrete and plaster. Aerospace may be more technology driven but smaller overall, people drive every day, but fly a few times per year.

The printer industry is like the auto industry. Everyone knows that the vertically integrated plants of the past are rare now. Frequently one plant makes engines that are used in several cars then ships them to assembly plants. Another plant makes headlights, and yet another the wiring loom. At the assembly plant the engine meets powertrain, alternator, radiator, lights, dashboard, door panels, seats, windows and trim coming from various other factories and other names in the business. So a thousand jobs gained in an assembly plant might translate to 10 thousand jobs in a regional economy.

Manufacturers are BMW, Chana, Chrysler, Daimler, FAW Ford, Fiat, GM, Honda, Hyundai, Tata (Jaguar), Mazda, Mitsubishi, Nissan, PSA, Renault, Suzuki, Toyota, Volvo and VW.

The manufacturers divide their efforts amongst brands, so BMW owns MINI and Rolls-Royce whilst Toyota operates as Daihatsu, Hino, Lexus and Scion. VW also operates as Audi, Bentley, Bugatti, Lamborghini, Scania, SEAT, and Å koda. etc, etc.

Manufacturers used to be vertically integrated. For instance Ford plants would make most of the parts for a car, and if they did not there would be a long relationship with the supplier. So for instance Henry Ford and Harvey Firestone were friends.

Ford also developed the franchise model of dealerships. Ford did not own the retail forecourts, but franchised its logo and brandname to be used by regional dealers.

There is still an element of vertical integration. Manufacturers run the risk that a critical supplier might fail and if it does the relationship becomes more visible. So Ford spun off Visteon in 2000 and they have built a list of other customers, and in turn spun off some of their unprofitable factories as Automotive Components Holdings. However that bit of financial engineering wasn't enough. Visteon UK went into administration in 2009 and Visteon filed for Chapter 11 reorganisation in May 2010 and emerged in October 2010. Ford effectively rescued Visteon.

Parts come from 3M(Automotive), Aisin Seiki, Alcoa, Alps Automotive (Alpine),American Axles, BASF, Bayer, Behr, BorgWarner, Bosch, Denso, Calsonic, Cooper Industries, Cummins, Dana, Delphi, Denso, DuPont, Eaton, Eberspacher, Evonik, Faurecia, Federal Mogul, Firestone, Freudenburg, Futaba, GE, GKN, Goodyear, Hella, Harman, Hella, Hitachi, Honeywell, Hutchinson, Illinois Tool, Johnson, Karmann, Lear, Lucas(TRW), Magna, Mahle, Mann+Hummel, Matushita, Michelin, Mitsubishi, NSG, NSK, Pioneer, Pirelli, Plastic Omnium, PPG, Rheinmetall, Rieter, Saint-Gobain, Siemens (VDO + OSRAM), SKF, Smiths Group, Sumitomo E GKN, Takata, Tenneco, ThyssenKrupp, TI Automotive, Toyoda Gosei,TRW Automotive (Northrop), TS Tech, Tyco Electronics, Valeo, Visteon, Webasto, Yasaki and Zeledyne - to name a few.

Some of these are household names, many not. People don't generally realize that Bosch not only makes household appliance and powertools, but is the world's largest supplier of automotive parts. Bosch invented the magneto - a high voltage generator that made spark plugs work (the logo is a little diagram of a magneto). But Bosch makes parts, not cars. Bosch does own Telex communications and makes IP-CCTV products.

Almost certainly Bosch, Denso, Delphi, Hitachi, Honeywell, Matsushita, Mitsubishi, Siemens, Sumitomo, ThyssenKrupp and Visteon could make cars - or indeed printers; they choose not to.

Almost every company listed has an interest in information technology. Some were once heavily involved like Honeywell. Alps made a popular line of thermal ribbon printer but the idea was largely displaced by inkjets and they no longer sell them outside Japan. Siemens made the first inkjet printers a decade before HP, Canon and Epson got involved.

Smith-Corona made printers until the 1980s but Smith group no longer seem to be involved with printers. TRW and relations were lead contractors on many US defence contracts.

Johnson Controls was originally a manufacturer of thermostats. They built building control systems originally using hydraulic pipes to work remote valves from a control centre. Johnson also a major manufacturer of batteries (Optima, Varta, LTH and Heliar brands) and of car seats, and building on that all kinds of interior trim.

Eaton make UPS, rivalling Schneider (APC).

Uses for Compatibles

We've outlined why technicians need compatible parts as well as manufacturers originals in an earlier essay .Briefly it's:

  • Competition  Even if you don't use compatible parts yourself it can be helpful if they exist because it helps keep the price of spares down. Some manufacturers think it is perfectly acceptable to charge half the price of a machine (or more) for a spare.
  • Environment  It makes sense to keep machines going forever if they are still capable of doing the job they were intended for. But the original manufacturers tend to end the supply of parts after 5 years or so.
Compatibles have a couple of bad points as well:
  • With no access to the manufacturers drawings, chemical formulations or patents a compatible maker is attempting to reverse engineer a part. That doesn't mean they can never make something better than the original, it's just less likely.
  • Compatible parts usually come in rather anonymous boxes - and so do most originals. The printer manufacturers themselves are not open about the origin of parts. Distributors, retailers and perhaps even manufacturers may sometimes sell compatible as original.

Printer manufacturers have a lifetime in mind for their product when they make it.   The word ‘lifetime’ is widely used as in ‘lifetime guarantee’ and not very meaningful in most cases. Long time computer-industry watchers know that when the top boss changes so do all sorts of policies.

Obsolescence

There is no inherent limit to the life of a printer.

A car rusts, the engines and gears are under stress and wear out. There are still 40 year old Hillman Imps and Morris Minors still on the road, but not many.

Aircraft are quite often 30 years old and sometimes more; providing the maintenance is good that isn't a problem (although counterfeit parts sometimes are).   A lot of industrial machinery lasts this sort of time, there are lathes and presses working that are 50 years old.

Computers are a bit different. A trusty old ‘ Apple ][ ’ will still work - but it's command line and Pascal editor will leave modern users dumbfounded. An easy to use graphics interface needs a processor a thousand times faster. So computers genuinely are obsolete after 5 years.   Printers less so.

Laser printers have changed since the first HP LaserJet in 1984. It could produce A4 pages with 300dpi graphics at 8 pages per minute. Modern printers are a bit smaller, faster and a lot cheaper to buy. So far as we know there aren't any first generation LaserJets still in use, they were big and heavy. People certainly are using LaserJet 4's from 1993. If a printer does what you want and you can still get parts why change?

Spare Part Policies

Manufacturers aren't obliged to supply spare parts at all. Most manufacturers of business equipment do supply some spares for more expensive equipment, both as a way to maintain their reputation and as an alternative profit centre. However manufacturers don't go on supplying parts indefinitely. HP has been quite good at supplying parts during their 30 years of printer making, perhaps reflecting their heritage as an engineering-oriented company providing laboratory equipment. Canon seems to have been more reluctant to supply end-users with parts, reflecting their heritage as a copier company with regional field service teams.

Manufacturers policies on parts vary. For some printers, small inkjets for instance, there simply are no parts except cartridges and possibly the power adaptor. The assumption is that a service technician could not repair the printer economically. Since the printer might be £40 and a call-out £50 a manufacturer might judge there to be no point making parts available for low cost printers.

It might be better if manufacturers were obliged to sell parts for a long period. That might drive more commonality of parts between successive models. Keeping useful things working might complement their WEEE obligations to dispose of equipment in a responsible manner.  Downsides would be cluttering toolrooms with ancient moulds and possibly slower progress. Paper feed mechanisms improved a bit between 1980 and 2010 so designs do have to change. Manufacturers would presumably hate such an imposition.

Spares policies can change. HP seem to have been having a rethink recently.   Some parts now come in HP branded boxes with an instruction sheet intended for end-user fitting. That may mean higher prices; there is always a downside.

Printer Lifecycle

When manufacturers withdraw parts people tend to think it's because they want to sell new printers. There are several reasons to think that, for instance:

Around the year 2000 manufacturers realised that they could put ‘killer-chips’ on cartridges. The printer tries to read a memory device on the cartridge and if it can't then it may refuse to work, or downgrade what it will do.

The advantage from the customer point of view might be that the cartridge can identify itself as correct for the printer, can have a built in sell-by date (toners can decay) and monitoring toner levels can be done by writing usage figures into the chip rather than by trying to measure it with magnetometers or suchlike.

The advantage from the manufacturers point of view is that customers are much more firmly tied to them. Killer chips make it more costly and difficult to refill cartridges or make compatibles. There is a quiet war going on where print manufacturers devise ways to encrypt their killer chips whilst cartridge manufacturers find ways to clone them (have a look at recycling magazines).

There are other reasons to withdraw parts.

Customers, rather than manufacturers, may actually drive obsolescence. At some point people think of buying a new printer anyway:

  • Performance is what manufacturers aim to deliver and customers presumably respond to. Looking at three printers over time:
    • Year, Model: Resolution,Speed , Memory
    • 1984, Laserjet: 300 dpi , 8 ppm. 128Kb
    • 2001, LJ-4100 : 1200 dpi , 24 ppm, 16Mb
    • 2010, LJ-P4014: 1200 dpi , 43 ppm, 96Mb
    Memory and processor power limited what the first LaserJet printer could do. In the 1990s resolution had reached 1200dpi but the printer slowed to handle it. The LJ4100 was one of the first where that didn't happen. Recent printers have 1200 dpi, some greyscale and the memory to cope with large bitmaps whilst printing at 20ppm or more. The newest printer gets its speed mainly from improved computer chips on a board called the formatter; mechanical things have improved a bit.
  • Slow print caused by the limited ability of the printer's formatter is the common problem in ten year old machines. The formatter is a local computer intended to offload tasks from the user's machine. The trouble is that on older printers 16MB of memory is not quite enough for the huge bitmaps some recent PDFs contain. There are remedies; check that the driver memory setting matches what is in the printer (print a config page to find out). Change from PostScript to PCL. The older generation of printers often have memory expansion slots, but DIMMs to fit them might be pricey.
  • Reliability of mechanical things does tend to fall over time. A succession of maintenance kits never really gets dirt out of crevasses and gearchains or dust out of the HV electronics and the laser scanner. Old printers can be cleaned up and made to work well, but it's a remanufacturing job. Buying new printers is probably an easier decision.
  • Contemporary styling is an issue. A lot of workplaces are partly functional but also intended to impress and new equipment might create a better image. Old printers tended to be cream colours, recently designers have used greys and even black.

Office printers are sold widely, but Fortune-Global-500 companies probably have a great influence on printer manufacturers because management teams meet to sign contracts. A lot of these companies do have an eye to style. Many lease equipment and have a rolling program of office equipment replacement. Basically if a few government department, aerospace companies, big banks and legal practices say they are still using the machines a manufacturer will think twice about ending supply of parts. The printer market beyond the Fortune 500 may be large but lacks influence.

There will be a demand curve for parts. Initially the manufacturer has one or more production runs of parts matching market forecasts for the printer model. As the model ceases production they may make a final run of spares. For things like rollers and pads which are forecastably going to be needed they might order a new batch at intervals.

Manufacturers probably find handling spares quite a costly exercise.

Cartridges are the key profit centre for most manufacturers. The simplest distribution is probably via traditional stationers, increasingly supplemented by supermarkets, but they sell few if any office laserprinter cartridges. Web sales are the other main outlet, there are a great many web-sites selling cartridges, usually getting them despatched by one of the distributors or by operations like ‘Staples’.

New printers sell through different channels to cartridges. Supermarkets often sell low-end inkjet models. The main outlet for business printers are the computer superstores such as PC-World. Other than for display items they in turn seem to operate from one or other of the large distributors price-feeds. Software houses and systems integrators are also suppliers; both will charge a bit more than other outlets, trading on the promise of convenience and support. Traditional copier dealers also sell printers but tend to be interested in the largest, which will warrant a leasing contract and scheduled maintenance.

Spare parts tend to go along a quite distinct logistics chain to that handling cartridges and new printers. In the past nearly all spares were sold to specialist service engineers - printers were something like ten times their price today so most people would call in a specialist. Spares go about four ways now: service engineers, systems integrators providing customer support, customer site IT managers and non-specialists who deal with IT matters on smaller sites.

Supermarkets and others who aim to actually have items in stock aren't equipped to deal with spare parts. A single printer model explodes to something like 300 parts of which about 100 are obtainable and perhaps 20 regularly required if you try to run a printer for a long time. Manufactures could standardise more across their range and that might make supermarket parts possible. It is often difficult to see why things like pickup rollers and pads need to differ between models - surely the same roller could be used in a great many models? However the fact remains that parts differ considerably.

People stacking the shelves in a supermarket aren't able to debate the merits of own-brand beans; how could they? Supermarket decisions are made in seconds. And yet supermarkets in the UK expect to make 10-15% on sales. They aren't likely to engage in printer spares where people are asking for some support, across a bigger range - for a lower margin.

There is a cost to having a stock bin, a mould in store, documentation on setup and assembly, and expertise on how things are done. Logistics departments measure these things. If the parts aren't meeting their costs they get withdrawn. We look at these costs in a bit more detail below.

Customers disposing of old printers at some point undermine the manufacturers ability to supply parts.

Anyone who wants to go on using these printers needs more than one source of parts. If there were one source they might charge as much as HP tends to (or more), and these are old printers so people don't really want to pay over £15 for a pad and over £200 for a maintenance kit.

Too many sources would also cause trouble. If there were too many sources, competition would drive prices down to the point where producers would abandon the market. And that might not happen in an orderly fashion leaving say three happily competing suppliers; instead everyone might quit.

A maker of compatible parts has to guess how many printers there are out there. The manufacturer knows, but they mostly aren't telling.

However there is no point us or anyone else making or stocking any of the parts unless we think there will still be a ‘critical mass’. There need to be enough printers out there for it to be worthwhile keeping a drawer with parts in. There is a cost to storage space, stocktaking, restocking and so forth.

Critical Parts

For a machine to remain viable all the critical parts have to remain available. These printers need about 20 critical parts which are:

C4129X cartridges   Life 10,000 pages. There are at least 5 manufacturers or remanufacturers other than HP/Canon including DataProducts, Xerox, Silver Reed and Longbow. The drums and other components will probably last two working cycles (1 refill) before needing a new drum. There are half a dozen parts needed to make cartridges - drum, scraper blade, doctor blade and PCRs seem to be available. There is nothing very special about the toner, its a common HP/Canon type. The cartridge refilling industry is large and tenacious so it would be a surprise if cartridges ceased to be available.


RG5-7061 Fuser   LJ-5100 fuser 220 Volt- Life about 150,000 pages. Refurbished fusers seem to be readily available in Autumn 2011. It's a film based fuser and the lubricant doesn't seem to be particularly critical. One UK distributor seems to have films. We'd suspect that the film that fit this fuser are probably used in several others, even if the fuser models do vary. Although other bits of fusers do crack with age its mainly the films that go, so they are the critical component. If the fuser films are available, recycling used cores is possible. The other components likely to fail are the heater, bushings, thermistor and thermal fuse - the last two are probably standard electronic products. A complete fuser would be quite complicated to make, there are several large plastic and metal components. We suspect there aren't enough of these printers for an alternative source of fusers like NuPro to make them. Because fusers are complicated and the parts will decay they could ultimately be a problem, but are not at present.

RG5-5460 Fuser   LJ-5200 fuser 220 Volt - same issues.

RG9-1542   Transfer Roller Assembly for LJ 5100. HP supply transfer rollers in maintenance kits and suggest changing them at 150,000 pages. Experience suggests that cleaning paper dust off with a soft brush also works - so the life could be 300,000 pages.

RG5-3579   Transfer Roller Assembly for LJ 5000. Metal shaft with foam roller. Although the part numbers are different for the LJ 5000, 5100 and 5200 we'd guess the manufacturing details for all foam precharge and transfer rollers used in mono printers might be quite similar. The main problem is probably to come up with a stable composition for a conductive foam rubber. Compatible transfer rollers have acquired a poor reputation. Cartridge recyclers address the same problems with the precharge rollers.

RB2-1820   Tray 1 Pickup Roller. Life probably about 150,000 pages. 2 parts.

RF5-3439   Tray 1 Pickup Pad. Life about 150,000 pages. 3 parts.

RB2-1821   Tray 2 Pickup Roller. Life probably about 150,000 pages. 2 parts

RG9-1485   Tray 2 Pad. Life about 150,000 pages. 5 parts including the spring.

Tray 1 rollers can be unused. We sometimes find users who didn't known that the pull-down door on the front of the machine is a tray. Other users put A4 in tray 1 and A3 in the cassette so tray 1 gets the most use. Most users buy a kit and replace both sets of rollers and pads at once.

The HP 5100 rollers are unusual because the wear pattern across their surface is even so wear on a roller isn't visually obvious - (and you can't just turn the rubber round a bit, which some allow).

The HP PartSurfer and Service Manual printer part lists run to 280 lines including whole assemblies like the motor and formatter. Most of those parts aren't worth making as spares. If they were wanted they could be sourced as “pulls” or as the industry tends to say “refurbished parts” - meaning they work and have been cleaned up.

Manufacturing Issues

Making printer parts isn't easy. We guess that setting up a laser printer production line will cost several million dollars and that something like a million might go on tooling to make the plastic parts .

Tooling up to make compatible parts won't cost anything like that, but to make the example RG9-1485 pad 5 components are needed and that will have a cost.

A professionally made plastic injection mould will cost something like $5,000. (We've been quoted £5000). OK, there may be ways to make injection moulds more cheeply - like the questionable part we've been looking at in part 1 .  However there does still seem to be an injection mould. There are flow lines and faults that show this part was injection moulded; it's not been made by a 3D printer, for instance.

Each roller and pad needs an injection mould for the ABS core of the thing. ABS seems to be used rather than polystyrene.


The rollers RB2-1821 and RB2-1820 don't look terribly difficult to make, although since the roller is moulded from the sides but goes flat on the paper, there isn't much allowance for a draft to help release of the part (measurement with digital calipers suggests it's slight).

Each roller also needs another mould for the rubber tyre. These could conceivably be cut from a long extruded strip. Each side of the strip has a beading. It isn't clear how the factory inserts the beading into the roller, it can be done with difficulty using a screwdriver shaft. Presumably some sort of jig that clamps the bead flat for insertion is used.


The metal pieces on the pad will almost certainly have been punch pressed from stainless steel, so that meant the expense of a punch and die. Since the punch and die would just be cut out of sheet metal themselves they wouldn't be so costly as injection moulds. There are problems because there are several bends and metal expands slightly when bent so there may have been a difficulty getting the exact measurement right.

The tray 2 pad needs a rubber pad, however that looks like its 10mm x 2mm thick and that could be an industry standard rubber strip. Neoprene strip looking very like this is available for under £2 per 10 meter roll (on E-Bay). The consistency should be as important as the dimensions and rolls on the web don't come with spec sheets giving friction coefficients.

The tray 1 pad is slightly simpler, just one metal trim. Annoyingly the rubber is slightly narrower, so perhaps it would be sourced separately. It could be shaved from a 10mm wide roll.

Making the rollers and pads for this printer needs four injection moulded plastic parts, four rubber parts with the roller tyres being moderately complicated and four metal parts. Setup is likely to be a significant cost and could be in the range twenty to thirty thousand dollars (pounds?) for moulds, punches and dies to make the dozen components. Historically HP and Canon have rarely used rollers from one printer in another, so each new model is a new set of tooling. The makers of the RG9-1485 compatible here got things wrong, so presumably they will have further expense of getting their tooling right. Regrettably we can't say they will lose market share because (as yet) they are anonymous. Correcting mistakes is a cost everyone has to bear sometimes.

Items like these cannot currently be made one-off; there is a minimum economic production quantity.

The rollers are just two parts, a rubber tyre on a plastic form. The plastic form probably could be provided by a 3D printer, outer dimensions are critical but that is all. The problem is to source and insert the rubber tyre. The pads pose more of a problem and have plastic, rubber and steel parts. Basically we have to make these things on injection moulding machines, metal presses and presumably with a bit of hand assembly. It is possible to order a short run of a few hundred pieces from some injection moulding plants but they prefer runs in thousands. Our guess is that a minimum production run might be 2,000 pieces, but the setup and mould making costs are then rather high. To get the unit cost down below $1 might need a run of 10,000.

The cost of materials in these parts is trivial - just a couple of pence. ABS may cost $2000+ per tonne but these rollers and pads weigh 10 grammes or less. So for an investment of something like the price of a mid-range car you can be the proud owner of a container load of rollers ‘worth’ a few pence each in material and nearly a dollar in setup costs. The real problem is, of course, how to peddle them to the market?

Marketing Rollers and Pads

Marketing is going to be the big cost.

The potential market is global and dispersed. If you are the manufacturer of compatible print rollers and pads it will be almost impossible for printer owners to beat a track to your door. Possibly they shouldn't; you are probably an injection moulding specialist who knows very little about printers or field-service.

First, people would have to find you through the blitz of entries in Google.

flp4it.com is interesting. Their Google entry reads ‘FLP Solutions Ltd. brings you the RG9-1485-000CN NEW from HP. We have over 10 years experience with delivering compelling solutions to help our customers ...’. Problem is this is a 14-year old product not ‘ NEW from HP ’. Compelling solution ? ... More like copperplate text that Google has bought into.

The search above was done on 22nd Sept. When I looked again (27th Sept) Google search result number had increased to 30,800 and 20,200 respectively. The order on the first two pages didn't seem much changed. Are the figures real? Does searching once prompt them to search harder? It might.

Google are trying to improve this kind of listing. They don't just look at the site contents. To maintain their listing position traders send them a daily xml data feed with information such as code, description, image, price, expiration_date, currency, brand, condition and product_type. Google changed to ‘brand’ from ‘manufacturer’ presumably because feedback showed that, whilst there is such a thing as manufacturing, it usually isn't done in the plant with the proud logo of a high-street name on the front. But the term brand is only a slight improvement - it implies a difference between rollers for the HP LJ 5000 and the Canon LBP-62X.

Spares Market

The computer spares market is being transformed, partly by the Web and search engines.

Spares Tradition

Computer spares have a long, complicated and expensive history. In the 1970s computers were very different from one another. Manufacturers like IBM kept supplies to themselves. Spares came from cannibalising scrap systems or from knowing who was the original equipment manufacturer - such as CDC who made most of the disk-drives or Dataproducts who made big printers. The market was relatively small and knowable. Sometimes suppliers could demand what they liked. One DEC printer made by Dataproducts needed a thermal fuse with a particular temperature. There were no alternative suppliers; they charged £240 for a part that was nearly matched by devices costing 35p. There was a choice: have printers fail prematurely, have the possibility of a fire or use the correctly rated part. The worst extortion we have seen for a mere printer part was a Honeywell photo printer with a failure prone switch. Just one third-party supplier could be found with the unusual switch; you could buy it for £550.

Purchase Prices Fall
Laser and inkjet printers became very popular. By the late 1980s as sales of PCs really took off there was a whole new market with manufacturers competing vigorously on printer sales price. Price and performance improved and prices slid over the years:
  • £3,150 in 1985 for a Canon LBP. (LJ-1)(PCW,MicroGeneral, July 1985)
  • £1,294 in 1991 for a Canon LBP-8-III. (PCW,Mediaware, April 1991)
  • £ 730 in 1991 for a Canon LBP4 (PCW,Mediaware, April 1991)
  • £ 972 in 1992 for a Canon LBP-8-III. (PCW,Novatech, June 1992)
  • £1,021 in 1992 for an HP LJ-III. (PCW,Novatech, June 1992)
  • £ 688 in 1992 for an HP LJ-IIIP. (PCW,Novatech, June 1992)
  • £ 711 in 1998 for an HP LJ-4000 (PCPro, Software Warehouse, Sept 1998)

Laser printer prices declined dramatically whilst performance improved significantly from 1984 when the first low-cost Laser and Inkjet printers became available. By the mid 1990's manufacturers likely sold at less than cost in the hope of making up the difference from cartridge sales.

The cheapest laser printers are less than £50 today, however those are cartridge burners. Business quality machines are still £350+.

Purchase prices have a big effect on spares. When the purchase price of a printer falls under $50 no service personnel are likely to look at any problem for anything other than altruistic reasons because the time they invest will not be repaid.

Disintermediation

There are something over 50 websites listing printer spares in the UK.   Some are like us, engineers who realised that customers had their own technicians and could do many jobs themselves. Some sites we think are just opportunists selling from distributor lists but they tend to have a recent web authoring package and use it to achieve a lively presentation with little content. About a dozen are "data-driven" - they take the price-feeds, add a percentage and put the result straight up. We all deal with the same group of distributors.

Distributors have their own websites. They aren't usually cheaper than operations like ourselves, their costs are high. For instance one of the distributors took 3 years to get their current website working and paid several hundred thousand pounds for it. It's a reasonable site but contents tend to be less informative than our own, they probably imported the data from their accounts system. Accounts systems have foreshortened descriptions and don't do much classification. Costs are reflected in their prices. They also tend to be part of large organisations so all their costs are high - support, sales order processing and invoicing for instance.

What the spares sites mainly do is function on a discount. We get something like 20% off the price the distributors sell at. To decide our online prices we get daily price-feeds from the distributors and turn those into a comparison list. Curiously the distributors disagree on pricing, sometimes by as much as 25% so on a few occasions one is half the price of another. The distributor who tends to have the lowest price also has a reputation for mistakes.

Some of our competitors don't seem to use a comparison, they just add 10% to 30% to their chosen distributor list and use that as their list. Everything is despatched by the distributor. In our case we hold stock of rollers but not usually of fusers.

All the spares websites would like to find the guys who actually make this stuff.

Individually and collectively we probably lack:
  • Purchasing power to get a containerload of stuff ordered and delivered to Tilbury. There are not dozens of orders per day for parts like the RG9-1485
  • Collective clout to get the goods quality checked and replaced if they are wrong

If you are making rollers you are probably running what is mainly an injection moulding plant. You probably don't want individual buyers - your primary interest is plastic, not printers. Furthermore selling products is quite expensive, there are payments to clear, transactions in foreign currency, customs forms and shippers to deal with. So typically you might have a minimum order of 500 parts, payment in advance.

If you are running an injection moulding plant you really need people to take large volumes, even though that means sacrificing a lot of the margin. There are about 30 printer distributors in the UK but two thirds of those only deal with new printers or cartridges, not parts.

HP have half a dozen parts distributors in the UK. One will only deal with people who turnover more than £250,000 in HP parts alone - that's a lot of parts. One is a broadline distributor who have a spares operation. There are three spares specialist distributors. One specialises in HP and reluctantly carries compatibles when there is no alternative. One is owned by a European operation who have an in-house compatible maker. There are probably three distributors who will consider taking bulk stock of compatible parts.

There might be another half dozen distributors across Europe. In the US perhaps the same.

Compatible parts makers face a struggle to market their wares.

Compatible Specialists

A few companies like Katun have been making spares for photocopiers for years, so they have some respect in the photocopier business, but they are unknown outside that market. The UK website wasn't working .

Clover Technologies are another global spares operation; 6,000 employees and $600 million annual revenue. They have the maintenance kits but not, it seems, the individual parts or a feed repair kit. They operate under several names, notably ‘ Dataproducts ’ a name they bought because that company had a long presence in the IT industry .

Both are quite large. Neither had a Wikipedia article when I looked, but that might be the Wikipedian perception of importance. Both probably commission their own parts from long established sources.

Suppliers like ‘Perfect Parts’ and ‘Microspareparts’ both seem to be distributor-captive names. Trust are a Dutch operation who seem to badge Gembird etc and have taken to producing some compatible printer spares.

There are a list of at least 40 Chinese outfits who claim to be making HP Parts. We treat these claims a bit sceptically. It's checkable to some extent. For the last couple of years HP have published a little list “HP Suppliers” and say

Below is an alphabetized listing of HP production suppliers. These suppliers represent more than 95% of HP's procurement expenditures for materials, manufacturing and assembly of HP's products all over the world. This list includes contract manufacturers, electronic manufacturing services providers, original design manufacturers, and commodity suppliers. HP is sharing this list with the intent of promoting transparency and progress in raising social and environmental standards in the electronics industry supply chain.

If a supplier isn't in that list then they are unlikely to have a direct relationship with HP. It is possible that they do have a relationship with Canon or with a subcontractor. We think there are far too many Chinese factories claiming to be making parts for HP for it all to be true.

Compatibles Market

Almost everyone in the printer market would welcome a reliable supplier of compatible parts. Manufacturers probably wouldn't.

Aftermarkets are almost inherently unstable.

Potential Market

It isn't clear how many of the Laserjet 5000 and LJ 5100 printers are out there. They were made for about 12 years. HP is the biggest presence in the printer market and Canon is second or third, so this will be one of the most common A3 printers. A3 printers were never terrifically popular, however. Banks and insurance companies will have got rid of their older printers; they might buy into the idea that chipped cartridges save money by always assuring the best print quality. Engineering companies are a bit more determined to keep things going and it's not just parsimony. If an old machine does a good job why change it?

One US site complains about compatibles using “cheap materials” but we don't think that is the problem. The defective metalwork wasn't a cheap material, it might be that the compatible manufacturer had cut out a manufacturing stage. However we think the real problem was that they just hadn't thought through how this part was suppose to work and if they tested the part at all they didn't test enough.

The plastic moulding certainly suggests the maker cut corners.