Dead HP 35

[teenix]

(10-10-2024 09:42 PM)krazyk Wrote:  I have been using 3 semi depleted AAA alkaline batteries that are showing around 1.3V each DC but I am aware that their current delivery capability can also be DEPLETED too. If I find out its THAT I will be very annoyed with myself.

I did think of that looking at the different batteries in the holder. I was thinking that sometimes just a single new battery can be added to partially flat ones and cause problems.

If you have a proper supply, try it out, maybe the batteries were the problem all along.

[HPing]

I have had great experience with Panasonic's Eneloop batteries.

The Eneloop Pro (black) batteries have more capacity but can "only" charge 500x. The regular Eneloop (white) batteries can be charged 2,100x. I have the white ones for a Walkman but can't remember why that was the better choice (maybe some audio foolery!)

I also have a basic Panasonic charger, which is nice quality and seems to work globally. The instructions stated batteries were to be charged in pairs only. I don't know if all of Panasonic chargers have that limitation.

I hope after all this work, the remaining problem is just a cable or battery or a bad soldier joint. Then you can concentrate on refurb best practices, whatever those may be.

[krazyk]

I recieved my white eneloop batteries today.....

And as expected, no change.

Total voltage delivered is 3.95 Volts DC

Vcc was LOWER this time at around 7.4 V.

The test point voltages were 5.4V and -10.9V so everything about 10% to 15% down.

I am guessing that the 35 was in such good condition physically because it has failed sometime in the mid to late 1970s and was put in safe keeping in a drawer.

I will fire up my Lab PSU just to be sure but I might replace the battery wiring just in case.

Beyond that this will require probing for clock signals and other signs of life.

[krazyk]

Just unpacked my in storage Lab Power Supply

Set the voltage to 3.8 Volts and set the current limit to 500mA and......

Suddenly Vcc jumped straight up to 8.15 Volts.

Vss still seems a little low at about 5.75 Volts and Vgg was 11.8 Volts

So finally the voltages seem okay eliminating the power supply components from being a problem.

Still no display and of course I now have a live readout of current flowing on the Lab PSU at 45mAmps.

Sorry Teenix but I should have unpacked this Lab Power Supply right from the start really as its just so useful now. Very occassionally when I handled the calculator the current jumped to around 300mA which was interesting. It may have been indicating a lack of continuity somewhere.

I have checked all of the gold connectors from the keyboard pcb but one appears slightly broken but well ABOVE the hole. I will try and post a picture tomorrow.

I am wondering whether it is an issue with power on reset as I had an Apple II motherboard fail to do that years ago. This might explain the 45 mA current usage. The engine of the calculator hasnt STARTED.

Kris

[krazyk]

Oops

Vgg was NEGATIVE - 11.8Volts I should have said.

[Geoff Quickfall]

Can you show us the power supply hookup to the calculator? At the calculator, please.

[krazyk]

   

[krazyk]

I tried resolving Phi 1 and Phi 2 clock signals too and I was simply getting a solid 5.6 volt signal with no resolvable oscillation and my DSO is a 200mhz unit.

I will try again tomorrow as it may be my ignorance again but I have succesfully resolved 1 mhz clock signals from Apple ][ motherboards with the same DSO using default trigger settings.

Could someone direct me to HP35 version 3 component diagrams as I would like to check the resistor values. Are any HP service documents available on the version 3 logic board?

Thanks again for everybodys interest and help but this is looking like super difficult at this stage.

Kris

[Geoff Quickfall]

Thanks for the pic. Now we know that the power supply and clip shorts are in place.

Check out this posting from 2010:

HP 35 chip description

In depth analysis

More!

Tony Duell hand drawn diagrams (incredible)

Waveforms on lipo power supply

And google more “HP 35 schematics”…

Cheers, Geoff

For me, now that I know the board is hooked up correctly to power and no display.

A) test continuity from prongs to board.

B) test continuity on all traces from component to component.

C) reflow solder joints at LED block. Reflow any grey powdery solder joints.

D) reflow any powdery, grey solder joints.

Thats to start.

Then voltage tests knowing all continuity tests are good.

Geoff

[krazyk]

Thanks for your reply Geoff

I have just tried to recapture Phi1 and Phi2 but this time at the KEYBOARD connector and there definitely seems to be something wrong.

This is Phi 1

   

This is Phi 2

   

Neither looks like any sort of clock signal I have seen and they just look like NOISE at voltages of 3.3V for Phi 1 and 3.7V for Phi 2 again measured at the keyboard connector.

It seems that the issue is likely up at the anode IC where a lot of solder appears grey and furry, particularly some of the VIAs there.

What do did you do about removing that plastic overlay covering the anode and cathode IC area?

Thanks again for your time

Kris

[krazyk]

The attached pic from previous post is Phi2 measured from the keyboard.

[teenix]

The easiest place to check the signals is by using the test port on the right side of the circuit board.

Use Tony Duells circuit diagram pdf as mentioned in Post #49 as a guide

cheers

Tony

[krazyk]

Thanks Tony

I have printed out the circuit diagrams and have tested for Phi1 and Phi2 clocks at both the test point to the right of the logic board and more recently from the ini and ini2 locations on the keyboard connectors.

In both cases I see voltage but no regular CLOCK pulses which seems consistent with a fault on the keyboard PCB associated with the ANODE IC and/or the grey furry solder from underneath where the battery box corrosion was.

I am about to resolder all the grey furry vias and solder points associated with the LED display and support ICs. Will do it gradually and will try not to cook the precious ICs. Was just concerned about the plastic overlay that HP have placed on the board.

Has anybody had this NO CLOCKS fault before here associated with this CLASSIC class of HP calculators?

Kris

[teenix]

Have a look at the clock driver IC, 1820 0855, (8 pins)
Check pins 1 and 4 for clock signals. These are connected to the Anode driver which generates the clock signals. If there are no signals here, then something might be wrong with the anode driver.

Image below shows the clock signals (yellow IN) (green x 2 = Phi1 and Phi2 OUT)

The IN signal might be a bit weak, I'm not sure how it will react to a probe.

It can be difficult to test the keyboard circuitry, so if you like, you can solder a temporary wire between the points (EDIT) marked blue which will short out the power switch. Be careful not to short adjacent tracks on the board. The CPU board can be connected to the keyboard, the battery wiring connected, and it should power up making it easier to look at the Anode/Cathode driver ICs. Be careful around the Cathode driver, it is a bit finicky when switching the LED coils, any upset may destroy it.

cheers

Tony

[ThomasF]

(10-14-2024 04:25 AM)teenix Wrote:  you can solder a temporary wire between the points marked (Green) which will short out the power switch.

I think that should be the Blue line (Green is indicating the clock signals).

Cheers,
Thomas

[AnnoyedOne]

(10-14-2024 04:25 AM)teenix Wrote:  Be careful around the Cathode driver, it is a bit finicky when switching the LED coils, any upset may destroy it.

"LED coil"??? Do you mean current limiting resistor? I've never seen an inductor ("coil") near LED's.

A1

[teenix]

The Classic LEDs are energized by small inductors that dump their charge into the LEDs to light them. It is a simple idea but is complicated in reality and the timing is fairly precise to get it working. HP did it this way as there is minimal power loss compared to using resistors.

Some details here near the start of this PDF document.

cheers

Tony

[ThomasF]

(10-14-2024 12:59 PM)AnnoyedOne Wrote:  

(10-14-2024 04:25 AM)teenix Wrote:  Be careful around the Cathode driver, it is a bit finicky when switching the LED coils, any upset may destroy it.

"LED coil"??? Do you mean current limiting resistor? I've never seen an inductor ("coil") near LED's.

A1

Then you should look closer to the classic internals ...
Just under the LED's there are two packages of inductor coils ...
   
Picture taken from here: Inside the 35
Also check Tonys excellent paper about the internals of the classics: Classic Notes.

Cheers,
Thomas

[AnnoyedOne]

(10-14-2024 01:15 PM)teenix Wrote:  The Classic LEDs are energized by small inductors that dump their charge into the LEDs to light them.

From p2 (3 in the PDF)

Quote:Further power savings were realised by charging small inductors and then discharging them into the LEDs to light them. This meant that the series resistors were not required and thus the power they consumed was eliminated. Another feature is that the displays are powered directly from the battery which means the power supply circuit does not have to be beefed up enough to supply the extra power.

Interesting! I've seen PWM (pulse width modulation) signals driving LED's to reduce power but not inductors.

Thanks for the information !

A1

PS: Looking at the PDF it looks like Tony has a DSO (digital storage oscilloscope). He must get paid well Nice traces though!

[krazyk]

This HP35 calculator is totally frustrating me.

   

This was the state of the cathode IC phi output solder from the main keyboard pcb from under the battery box region.