Unusual HP-21 ?

[jch]

Hi,

I've just received my first red LED's HP!
But unfortunately, it exhibits the problem described in this thread: http://www.hpmuseum.org/forum/thread-264...ight=hp-21

I took a look at the internals and it seems the PCB is not the same as usual.
It has got an AMD 1820-1983 (power-on reset, battery low comparator) and a dedicated power supply converter.

Did someone already encounter this version?

I have to check if there is some noise on the power supply, and if unsuccessful it will become a host for Panamatik's Woodstock low power module!



s/n: 1710S23250

Jean-Christophe.

[jch]

Indeed Vss exhibits a >10% ripple at about 70 kHz.

I have tested the power module after removing it from the main PCB. With a load of about 30 mA on Vss the ripple is still present.

May this quite high ripple prevent the operation of the ACT?
Or is it the ACT that needs to be replaced?

[jebem]

I would just replace all the electrolytic capacitors.
One of them is dry and not able to filter the rectified output from the diodes.
Or remove them, one by one, and measure them for capacitance and internal impedance if you have the right equipment.

By the way, i really like your oscilloscope!
In the 70's it was the best brand of lab equipment (just expressing my opinion, based on real experience with other competitor brands of the time).

[PANAMATIK]

(11-11-2017 09:08 PM)jch Wrote:  Did someone already encounter this version?

I have to check if there is some noise on the power supply, and if unsuccessful it will become a host for Panamatik's Woodstock low power module!

Jean-Christophe.

This is exactly the HP-21 board version which I used for one of my low power prototypes. You will need some extra work to install the LP module, because the display driver chip distance is different to normal boards (see LP manual "Replacement").

You can try to apply 6V and -12 V directly from a (ripple free) lab power supply to pin 1 and 2 of the ACT and check if it works. This would proof that the switching power supply is causing the problem.

Bernhard

[jch]

(11-25-2017 11:00 PM)jebem Wrote:  I would just replace all the electrolytic capacitors.
One of them is dry and not able to filter the rectified output from the diodes.
Or remove them, one by one, and measure them for capacitance and internal impedance if you have the right equipment.

Yes, replacing the capacitors was my first idea but looking at the schema (different from the HP-25 documented by Jacques Laporte), there are only two 'prone to dry' electrolytic capacitors:
- One at the Vbat input, which also exhibits some ripple. I already replaced it but to no avail, still the same ripple.
- The other capacitor filters Vgg which is quite clean but is the next candidate for replacement.
Vss and Vdisp are filtered by small tantalum capacitors, would they classify as electrolytic capacitors which may dry?

(11-25-2017 11:00 PM)jebem Wrote:  By the way, i really like your oscilloscope!
In the 70's it was the best brand of lab equipment (just expressing my opinion, based on real experience with other competitor brands of the time).

I'm sorry it's no mine , I had to borrow it from a friend where it was lying idle.
I agree with you, in my first job I had to maintain lab equipment and I made use of this model and repaired a few other. It is really a pleasure to use it after so many years.

[jch]

(11-26-2017 07:44 AM)PANAMATIK Wrote:  This is exactly the HP-21 board version which I used for one of my low power prototypes. You will need some extra work to install the LP module, because the display driver chip distance is different to normal boards. (see LP manual "Replacement").

Yes, I was thinking about that, but how is the LP module dependant on the power supply?
And I will get into that dilemma too.
I'm eagerly waiting for this little gem and I hope I'll be able to perform a successful transplant.

(11-26-2017 07:44 AM)PANAMATIK Wrote:  You can try to apply 6V and -12 V directly from a (ripple free) lab power supply to pin 1 and 2 of the ACT and check if it works. This would proof that the switching power supply is causing the problem.

However, the power module supports also the inductance and capacitor for the oscillator and the POR signal generation (derived from Vss through another capacitor).
Before removing the power supply module, the clocks were present, but I would bet this is not enough to report that the ACT is healthy?

[jebem]

(11-26-2017 08:12 PM)jch Wrote:  Vss and Vdisp are filtered by small tantalum capacitors, would they classify as electrolytic capacitors which may dry?

Tantalum caps usually fails catastrophically by going into a short circuit, which is not the case here. However, I have seen these caps presenting leakage, in other words, they act more like resistors than filtering components.

When i see ripple above the acceptable value and there are just a few low cost caps to look for, and i can't find any other obvious causes, i don't think twice, and to save time, i just replace them.
Obvious causes would be:
- Any component in parallel with the output that is draining current above the power supply specification (a defective cap is the usual suspect when the power supply was disconnected from its target load, as you did);
- Any defective power supply component. I would start looking to the rectifier diodes, they can present reverse current leakage above the acceptable. Again, when in doubt, i just replace the diodes and move on.

Edit:
I see a lot of pcb traces corroded by battery electrolytes.
Did you try to remove the capacitors?
I would not be surprised if one or more solder joints in the power supply were affected too.
That would isolate one or more components.
It is easy to spot such solder joints: It is very hard to unsolder them.

[jch]

Much later...

Replacing every component of the PSU daughter board, the one after the other, without spare parts readily available, seemed rather unworkable. I took apart the whole thing. Simulation with LTspice confirmed that the ripple was due to tantalum caps' ESR worst case, but it reads quite lower with the module rebuild on a breadboard with all new components.

Alas, still unsuccessful...

As Jebem has noticed, some pcb traces were badly damaged by batteries electrolyte.
And yes, desoldering corroded solder joints was tricky!
Cleaning, desoldering ACT and POR/Bat, re-cleaning, checking pcb traces and resoldering didn't help.

Jacques Laporte's pages and Jim Johnson's "HHC 2013 - HP29C Diagnosis and restoration presentation" are great resources and I was able to check some ACT's waveforms.
Clock phases 1 & 2 and ISA seems Ok but POR (ACT Pin 3) exhibits a square signal. So, once again, desoldering POR/Bat then connecting to pin A2 as on Jacques' schema, but still without success.

I feel like putting everything back in the drawer until I can put my hands on another HP-21 or not being able to withstand the call for an LP module...