Power supply fix for Canoscan 8800f

Question

This is essentially a revival of an old thread (from Nov 2016), but I am a newbie and am not able to make comments or ask further questions on it...

Canon no longer makes CanoScan 8800F AC adapters, help needed to replicate its functionality

In a nutshell - the Canon Canoscan 8800f scanner has a unique external power supply unit which supplies 32V dc, plus has a second 32V dc line which has a 1.5k resistance across it for 3 seconds. These Canon power supplies are almost impossible to get, and somebody had the good idea of taking a particular HP printer power supply, which also gives 32V dc, and inserting a circuit on this second supply line which performs the function of giving it the 1.5k resistance initially, then removing it with a timer, as this scanner requires.

In this previous thread, as linked above, two users here kindly offered designs for this delay/resistor-bypass circuit. At first I tried the second one, because it was more simple, submitted by 'Whit3rd'. This circuit functions, but it kicks in too soon, almost instantly. The scanner only responds and switches on if the delay is 2-3 seconds.

Looking now at the first circuit design, offered by 'ThreePhaseEel', it has more components, and I am about to order the components I haven't got to make it but...

I've three questions:

A: With Whit3rd's version, which components can I experiment with to increase the delay of it's response? Any suggestions of configurations to alter this?

B: With ThreePhaseEel's version, again, which components would most impact the delay, so I could substitute parts to get the right result?

C: Again also with ThreePhaseEel's design: could somebody please clarify/confirm the choice of capacitor in the design: ThreePhaseEel writes that '...the timing capacitor should be a stable, low-leakage type such as a polyester film or a good grade of tantalum cap...' - yet the cap is listed as 100uF which seems large for a Tantalum and is near-impossible to get as a Polyester Film... Is this correct? I have electrolytic caps this size. And what voltage for this cap? 50V seeing as the circuit is 32V?

Thanks in advance for any responses. I am not trained in electronic engineering, more an intermediate-level tinkerer. I can easily make these circuits, as designed by other people, but I don't claim to understand them to the point that I can modify or redesign them.

Answer 1

ThreePhaseEel's design and Whit3rd's version appear to only be theoretical. Considering the lack of information on why this 'short the resistor after a few seconds' trick is needed, I would stick with the relay based time delay circuit since it is known to work.

But first I would try just wiring up a 1k5 resistor with a mechanical switch across it. From this I would find out whether the technique works at all, and what time delay is required. And if the timing can be reliably done by hand then why bother with a more complex circuit?

Answer 2

Nowadays I would not mess with analog timing circuits anymore for hobby tasks. An ATtiny85 microcontroller board with USB interface is available for 2-3 Euros, which is the monetary equivalent of like a few minutes of my working time (which one should always consider for private projects as well).

So if you're not particularly interested in learning an analog electronics lesson but just want to get the job done, order such a board (+ voltage regulator for 5 Volts + some auxilliary parts), connect a suitable logic-level MOSFET gate to one of the ATtiny's ports and write a short code snippet for switching the MOSFET (essentially this is hardly more than the arduino blink sketch). Voila, switching timing with at least internal oscillator precision. If you need higher timing precision, just order a cheap arduino clone with a quartz oscillator instead (~ 4-6 Euros, still within said opportunity cost).

Changing the timing then is a matter of typing two numbers into your code and hitting the compile/upload buttons with your mouse. No endless cycles of soldering to stripboard or something until you get the values right.

Answer 3

I had luck with this scanner today sending 32 to the right pin and running 32 through a 2.5 k Ohm pot, when rotated to the full 2.5 the scanner booted and functioned, did have to hold the power button down for a few seconds to convince it to boot though