Author Topic: Keithley 2306 Dual Battery Simulator Repair (maybe) (Read 6857 times)

Smokey · « **Reply #25 on:** July 23, 2022, 02:53:43 am »

Quote from: AVGresponding on July 22, 2022, 05:41:49 am

...Have you checked the input ripple voltage? If it's reading close on DC, it might be falling under the dropout voltage at the bottom of ripple voltage. Though the output you measured looks good to me; I'd be surprised for anything designed to work using a standard 5V linear regulator like this to stop working at 4.94V if it's not faulty in some way.

About 15mv ripple on the +6.5VD rail. There is some +/- 50mV high frequency stuff coming from the switching regulator frequency that's generating the voltage, but I wouldn't say it's excessive.

U112 is the switching regulator that's driving the +6.5VD rail. It's listed as "ADJ VOLTAGE REG STEP DOWN" in the component list on the data sheet. Thanks Keithley.
The part is actually a LM2675:
https://www.ti.com/product/LM2675
https://www.ti.com/lit/ds/symlink/lm2675.pdf

The output from the regulator is essentially a 260kHz 0-12V square wave into the inductor/cap.

R124 = 2.55K
R125 = 11K

So assuming R124=R1 and R125=R2, The output voltage is set to:
(((R2/R1) + 1) * 1.21) = 6.429V, which is pretty much right on.

NOTE: Since this looks good, no need to get any other measurements on this. It is what it is.

Smokey · « **Reply #26 on:** July 23, 2022, 03:30:03 am »

As a test, I bumped R124 to 2.4k.
...Which increased the measured input voltage of the LM2940CT to 6.68V.
The measured output of the LM2940CT is still super solid at 4.95V, and the 2308 supply still won't consistently power up with just the internal regulators.

So... It's not looking like it's an issue with the input to the LDO powering the +5VD rail.

AVGresponding · « **Reply #27 on:** July 23, 2022, 09:01:37 am »

Time to start looking at the control logic then... IC decoupling caps, reset signals etc...

Smokey · « **Reply #28 on:** July 24, 2022, 07:09:21 pm »

Filling in some part numbers for "Model 2306 digital board":

U132 = "INTEGRATED CIRCUIT " = LM2937 = "500-mA, 26-V, high-PSRR, low-dropout voltage regulator"
https://www.ti.com/lit/ds/symlink/lm2937.pdf

U135,U136,U235,U236 = "INTEGRATED CIRCUIT" = LM7121 = "235-MHz Tiny Low Power Voltage Feedback Amplifier"
https://www.ti.com/lit/ds/symlink/lm7121.pdf

On the "Model 2306 display board parts list"
U601 = "UNDERVOLTAGE SENSE CIRCUIT" = MC34064
https://www.onsemi.com/pdf/datasheet/mc34064-d.pdf

Smokey · « **Reply #29 on:** July 24, 2022, 07:55:18 pm »

Checking some processor lines:

This is all in the "failed to power up" state.
{
Checking (not)RESET, (not)HALT, and (not)BERR line on the MC68331CPV16 Main processor...

Reset line goes high about 525ms after 5V rail comes up. Stays high.
HALT line goes high at the same time as 5V rail. Stays high.
BERR (bus error) is also high at power up. Stays high.

Clock out looks good at 25MHz
Freeze line stays low.
}
Nothing strange here.

Powering the system with an external 5V (successful power up)
{
Same behavior. No changes to the above signals
}

That all looks good.

Smokey · « **Reply #30 on:** July 24, 2022, 08:45:55 pm »

Note: As I said above, the main digital board does not consistently power up regardless of whether the display board is plugged in or not. Also I have captures of the display board's serial communications in both working and not working conditions, so I'm not saying this is causing the issues with the digital board... or is it?

.....

The MC34064 undervoltage sensing circuit on the display board is acting strangely..
From this plot in the datasheet...

The output is supposed to snap on around 4.5V input. It's not exactly doing that.
At about 4.4V input it starts to turn on the output, but it's output is only 1.5V.
4.5V in = 2.2V out.
4.6V in = 2.75V out.
4.7V in = 3.3V out
4.8V in = 4.0V out
4.9V in = 4.9V out
5.0V in = 5.0V out.

That isn't even close to what the plot says it should be doing. The IC itself doesn't look damaged and it's not getting super hot or anything.

So that may explain why the display only turns on at the full 5V input....
but... at least according to the parts list, there is no voltage monitor IC on the digital board and the main digital board does not appear to rely on the display board for anything since it will power up when it's totally disconnected.....

Kean · « **Reply #31 on:** July 24, 2022, 09:47:50 pm »

The MC34064 has an open collector output and so depends on a pull up resistor.
Maybe the MC34064 is damaged, or more likely something else is loading down the reset line.

Smokey · « **Reply #32 on:** July 24, 2022, 10:05:33 pm »

The pullup on the "Model 2306 display board parts list"
R612 = RES, 4.75K, 1%, 100MW, THICK FILM

If the comparator in the MC34064 was acting correctly, I would expect the open collector to pull hard to ground until it hits the threshold and the output snaps on. But its got this strange ramp that's under spec before the turn on. Once the input gets to 4.9V it's full on like it's supposed to. If there was excessive load, I would expect a voltage droop here too, but it's solid above 4.9V.

But even at the top range of the input specifications in the data sheet, the highest it should get before it snaps on is 4.8V. So even if we ignore the ramp voltage, it is turning full on too high.

Kean · « **Reply #33 on:** July 25, 2022, 03:25:45 am »

So two possibilities I see
1. The MC34064 output is not changing state cleanly at 4.5V, which seems unlikely as it has a simple comparator driven output with hysteresis
2. The MC34064 output is changing state at around 4.5V, but between 4.5V and 4.9V other circuitry loads the signal more than the pull-up resistor can handle

The fact you see a change at around 4.4-4.5V makes me think the latter is the case. You could isolate the output pin of the MC34064 and see what happens to the reset signal voltage with just the 4k75 pull-up as you vary the supply from 4V to 5V. Or just swap out the MC34064 to eliminate that as the cause.

Smokey · « **Reply #34 on:** July 26, 2022, 07:45:46 pm »

Success. (so far.....)

I ended up doing the mod I described above, where I replaced the original LM2940CT generating the +5VD rail with a MIC29302WT adjustable regulator.

I set the MIC29302WT output to 5.10V, and the whole supply works fine now. (The original LM2940CT had an output spec of +4.85 to +5.15, so I don't see this causing any issues with the other parts of the system).
External VFD display turns on with good brightness (The 5.1V pushes the display MC34064 voltage monitor into the region that's full on, so that problem goes away as well).
Digital board powers up every time. I tested the output of channel1 and it's still accurate. It's been put back together running in it's home on the bench all morning with no issues.

In my desire to be done with this project, I didn't run some additional experiments that I probably should have:
1) I didn't use the pot to drop the output voltage down to the 4.95V that the original LM2940CT supply was making when the unit wouldn't turn on consistently.
2) I didn't just replace the LM2940CT with another LM2940CT and see if that fixed it as well.
3) I didn't replace the MC34064 on the display board (even though I'm suspecting that the 5V reg was causing this strange issues as well somehow).

If I had to guess, I think the problem was most likely just a faulty LM2940CT (even though the output looked fine on the scope). If anyone else has a 2302/2306/2308 that isn't powering up, I would start by just replacing the LM2940CT.

Smokey · « **Reply #35 on:** July 27, 2022, 05:41:53 am »

One more post here to thank everyone that posted ideas, and helped troubleshoot. You guys are the best!

Smokey · « **Reply #36 on:** October 15, 2022, 02:49:30 am »

Grrr..... Saga continues... This power supply has been acting up again.
It's been need to warm up for a couple minutes for the internal display to work. Once it's warm it will work fine when you power cycle it... at least it was doing that.

Now the internal display won't come on even after it's been running..... The external display still works though so I know the unit is powered up and working.

I'm going to replace the MC34064 voltage monitor chip on the internal display board, and I'll see if that fixes it.
Parts ordered.

ottobad · « **Reply #37 on:** July 14, 2023, 01:50:47 pm »

Hello, I have a Keithley 2306 and it suddenly appeared INITIALIZING ... and that's all it does. I took it apart and measured internal source , its source outputs 12VDC, after that the LM2940 outputs 5VDC, and then if I measure TP 101,102,103...nothing. On the CR102, rectifier bridge, I have nothing on the alternative and nothing on the DC, as well as the next module. I measured on SRAM and 2.5VDC comes in... what should I measure or where should I start?
Thank you and sorry if I disturbed you.

Kean · « **Reply #38 on:** July 14, 2023, 02:23:12 pm »

Quote from: ottobad on July 14, 2023, 01:50:47 pm

Hello, I have a Keithley 2306 and it suddenly appeared INITIALIZING ... and that's all it does. I took it apart and measured internal source , its source outputs 12VDC, after that the LM2940 outputs 5VDC, and then if I measure TP 101,102,103...nothing. On the CR102, rectifier bridge, I have nothing on the alternative and nothing on the DC, as well as the next module. I measured on SRAM and 2.5VDC comes in... what should I measure or where should I start?
Thank you and sorry if I disturbed you.

I think you need to enable the channel before voltages will appear on those test points. If it is stuck on INITIALIZING then I guess you cannot do that.
If the main 5V voltage looks OK, then per the above post #10 the SRAM chip could be the culprit. You mention 2.5V on the SRAM, but I think it should be powered from 5V rail. What pin did you measure 2.5V?
I bought mine on eBay knowing it may have similar issues, but so far it has been fine after I recalibrated it.

ottobad · « **Reply #39 on:** July 14, 2023, 04:59:55 pm »

It stuck on INITIALIZING, also no service manual on internet, anyone have it here? I measure SRAM chip on pin 32Vcc and pin 16, i have around 2.5Vdc. Do you know who supply voltage to sram chip?
I will order from Ebay the SRAM chip....Being a RAM, does something need to be written, copied from the old one?
Thank you for reply.

Kean · « **Reply #40 on:** July 14, 2023, 05:23:50 pm »

On the SRAM pins 16 and 32 are GND and Vcc and my unit reads 4.95V. There is a capacitor C186 across the SRAM supply rails near pin 32.
The SRAM is not battery backed in the K2306 - no battery I can see and no voltage to the SRAM when powered off, so no need to copy any data.

The R1LP0408C SRAM needs 5V +/-10% so 2.5V is bad. If you really have 2.5V then I suspect that may not be caused by a RAM chip fault if you measure 5V elsewhere.
There are a couple of different voltages in the digital section, and a bunch more isolated supplies in the analog area. I will make notes of some measurements for you to compare.

Kean · « **Reply #41 on:** July 14, 2023, 05:45:49 pm »

OK there are three isolated supply sections on the K2306 main PCB.
I've listed them below along with measurements from my K2306 with the top PCB removed (so no load on analog supplies).

Digital supply "D" (test points near fan) - referenced to GNDD
TP105 Vin = +11.97V
TP106 GNDD
TP109 +6.5VD = +6.47V
TP110 +5VD = +4.96V (this connects to the SRAM Vdd pin 32)

Also there is a second 5V regulator in the digital section at U132 with pin 1 = +6.42V (from +6.5VD via filter) and pin 3 = +4.96V. Not sure what this supplies.

Analog supply "A" - referenced to GNDA
TP100 +5VA = +4.96V
TP101 +15VA = +14.87V
TP102 GNDA
TP103 -15VA = -15.01V
TP104 -5VA = -9.50V
TP111 +25VA = +37.18V

Analog supply "B" - referenced to GNDB
TP200 +5VB = +4.97V
TP201 +15VB = +14.96V
TP202 GNDB
TP203 -15VB = -15.08V
TP204 -5VB = -9.52V
TP112 +25VB = +37.42V

* I was wrong about needing to enable channels to read all measurements - you just need to use the correct GND pin as they are isolated supplies.

Kean · « **Reply #42 on:** July 14, 2023, 06:15:13 pm »

Before putting my K2306 back together, I took a little look around the main CPU (MC68331)

U129 and U133 are 24LC16B I2C EEPROMs which probably store the calibration data - much better than battery backed SRAM, but not infallible. When I got my unit the calibration was messed up, but that could have been someone playing around before me rather than any kind of failure.

U132 might be a separate 5V regulator for the remote display connector so if it is shorted it doesn't take out the main 5V rail. Or it could be a clean supply for the nearby analog components: 2 x LM393, MAX391, AD706. I don't know what they are used for.

There are 2 x ADM232 which will be for the remote display comms, a ULN2003 for the relay outputs, and some opto-isolators for comms to each of the isolated analog channels on the top PCB.

ottobad · « **Reply #43 on:** July 18, 2023, 07:48:40 am »

Hello,

I have the :
TP105 Vin = +11.8V
TP106 GNDD
TP109 +6.5VD = +6.47V
TP110 +5VD = +5V
On U132 in 6.4V out 5VDC but on SRAM pin 16-32 IS still 2.5VDC...Is ok If I take out the memory and mesure again if 5 Vdc is present on pin16-32...maybe is a faulty memory?
On this...I have 0 insted of value you put

Analog supply "A" - referenced to GNDA
TP100 +5VA = +0V
TP101 +15VA = +0V
TP102 GNDA
TP103 -15VA = -0V
TP104 -5VA = -0V
TP111 +25VA = +0V

Analog supply "B" - referenced to GNDB
TP200 +5VB = +0V
TP201 +15VB = +0V
TP202 GNDB
TP203 -15VB = -0V
TP204 -5VB = -0V
TP112 +25VB = +0V
Thank you

Kean · « **Reply #44 on:** July 18, 2023, 08:25:20 am »

Quote from: ottobad on July 18, 2023, 07:48:40 am

On U132 in 6.4V out 5VDC but on SRAM pin 16-32 IS still 2.5VDC...Is ok If I take out the memory and mesure again if 5 Vdc is present on pin16-32...maybe is a faulty memory?

As you measured 5V OK at TP110, then I suggest measuring the supply pins of a few other components in the digital section other than the SRAM.
Did you try measuring across capacitor C186? Other points to try measure would be between GNDD and VCC on the various 74-series logic ICs, or ther I2C EEPROMs U129, U133.

You could also measure the resistance between the supply pins on various ICs in the digital section and pin 32 of the SRAM (with the power off of course) to see if there is some broken connection.

I don't believe there is a service manual or schematic available. One user here (picburner) created a partial schematic of the digital to analog opto-isolation section at https://www.eevblog.com/forum/repair/keithley-2306-dual-psu-battery-simulator-repair/msg632229/#msg632229 and three people in that thread fixed (or partly fixed) their unit by replacing the SRAM IC.

Quote from: ottobad on July 18, 2023, 07:48:40 am

On this...I have 0 insted of value you put

Just to make sure... when you measure these analog section supplies... you do have your DMM black probe on GNDA or GNDB and not GNDD?

I will open mine up again and see if I can identify any other suggestions.

ottobad · « **Reply #45 on:** July 18, 2023, 09:04:25 am »

So, C186 ( shorted ) is 4.98Vdc and now SRMA is at 5V BUT STILL IN INITIALISING ( don't know the value of C186 so is out from the board )
TP110 is ok, 5Vdc.
I will buy the SRAm and after recieve, will try again.

Kean · « **Reply #46 on:** July 18, 2023, 09:40:15 am »

OK. C186 is 100nF 20% 50V Ceramic per the list of parts in the manual. I don't think the exact specifications are critical.
Let us know how you go after replacing the SRAM. i hope that fixes it.

For your two isolated analog power supplies that read 0V on the outputs, check the input fuses F100 (marked FH100/FH101) and F200 (marked FH200/FH201)
They should read approx 12V (relative to the non-isolated digital GNDD) on both sides of the fuses, or should at least not be open circuit.

Smokey · « **Reply #47 on:** July 18, 2023, 09:18:54 pm »

Hey guys. Sorry for not getting back sooner on this. I actually never fully updated what I discovered on my unit.

Firstly I have about 10 of the SRAM chips that I bought when I was first doing my investigation, but never used because that wasn't a problem I actually had.
https://www.eevblog.com/forum/buysellwanted/fs-r1lp0408csb-5sc-4mb-sram/

Before you go swapping that, here is what I did that actually fixed my unit (at least made it reliable enough that it appears to work without any of the errors I was originally having).

Firstly, if I remember correctly, the "Initializing" on the display is actually coming from the display board. So it's not some informative error message that the main processor is trying to tell you, it's just what the display board shows at power up when nothing else is talking to it.

Ok, so what I determined was the actual problem with my unit was a bad cap on the DISPLAY BOARD that was pulling down 5V for the rest of the system. Unfortunately, the cap is between the display and the PCB which are soldered together with a lot of pins. So instead of disassembling that, I just stuck a TON of extra capacitance on the top side of the display board 5V rails. There are at least two unpopulated 5V cap footprints on that display board, so I stuck two 100uF 5V tantalums which I had around on those pads. That gave the board enough capacitance that it no longer pulled down the 5V rail, and everything was happy after that.

Hopefully that helps.

ottobad · « **Reply #48 on:** July 19, 2023, 09:55:38 am »

H :-//ello, thanks for the advice, I disassembled the display module, checked all the capacitors, they are good. I checked 5Vdc, all good. All that remains is to get the SRAM and change it... maybe that will fix it


EEVblog Main Site	EEVblog on Youtube	EEVblog on Twitter	EEVblog on Facebook	EEVblog on Odysee

Author Topic: Keithley 2306 Dual Battery Simulator Repair (maybe) (Read 6857 times)

Share me