11801, 11801C and CSA803A repair projects


Reginald Beardsley
 

This started as a conversation among @oculus, Jerry Hancock and I. Jerry got an 11801C, @oculus got a CSA803A and I've got a 2nd 11801 which Jerry in a fit of foolishness sold to me ;-) along with an SD-26 with a bad channel that turned out worked perfectly the 2nd time I tested it before shipping it back to Jerry. But it's still causing Jerry trouble. Fortunately, I took photos and can prove it *really* did work correctly in the far left slot of 11801 #1. When I got it and tested it, I wrote on it in pencil, so I know it's the one I got from Jerry and not one of the set of 4x I bought for $$$.

As I was finishing this I realized we really should archive the work on the list. The previous emails don't amount to anything, but at this point it's worth archiving.

Have Fun!
Reg

-----------------------------------------------------------------------------------------------------------------------------------

I've got 11801 #2 on the bench. There is no calibrator output. Interestingly it passes all the diagnostics but they don't seem to run automatically. However, there is a way to jumper those off which I'll need to look up, pull the correct board and check.

I just discovered that there are additional diagnostics windows for each of the blocks across the top of the Utility screen which vary with the subsystem selected on the LHS.

On 11801 #1 setting the trigger to "internal" starts the calibrator verified using a DSO. But it does not on #2 which explains why I couldn't get autoset to work on #2. IIRC I tried running #2 in the stack in the dining room and got error messages from the diagnostics.

The SD-26 I sent back to Jerry, when it came didn't work properly on one channel verified on 11801 #1. Before I sent it back to Jerry I tested it on #1 again and it worked. But Jerry is having problems with it.

I have now verified that the calibrator is getting an input signal, but not producing output. I don't yet know if it is getting power. I hope it's not. That <20 ps source would be pure unobtainium and unrepairable without a wirebonder and an almost unobtainum tunnel diode. One could build a replacement using the same Maxim chip that Leo uses which has a rise time spec of 21 ps. It might well be a bit quicker with a smaller step.

The input to the calibrator is a square wave so it should be possible to make a replacement for the crappy 250 ps rise time calibrator in the 11801C. Should be a very simple board and *very* popular with 11801C owners. I don't know when they changed from the marvelous <20 ps pulser to the 250 ps pulser, but learning that killed my interest in getting a C unless there is a way to put a better calibrator in it.

11801 #1 is equipped with an SD-22, SD-24, SD-26 and SD-32 set and now occupies a place of honor on my very small bench.

On Saturday, March 27, 2021, 01:57:11 PM CDT, Sean Turner <sdturne@q.com> wrote:

#yiv6493947978 #yiv6493947978 -- _filtered {} _filtered {} _filtered {}#yiv6493947978 #yiv6493947978 p.yiv6493947978MsoNormal, #yiv6493947978 li.yiv6493947978MsoNormal, #yiv6493947978 div.yiv6493947978MsoNormal {margin:0in;font-size:11.0pt;font-family:sans-serif;}#yiv6493947978 p.yiv6493947978msonormal, #yiv6493947978 li.yiv6493947978msonormal, #yiv6493947978 div.yiv6493947978msonormal {margin-right:0in;margin-left:0in;font-size:11.0pt;font-family:sans-serif;}#yiv6493947978 p.yiv6493947978msonormal3, #yiv6493947978 li.yiv6493947978msonormal3, #yiv6493947978 div.yiv6493947978msonormal3 {margin-right:0in;margin-left:0in;font-size:11.0pt;font-family:sans-serif;}#yiv6493947978 p.yiv6493947978msonormal41, #yiv6493947978 li.yiv6493947978msonormal41, #yiv6493947978 div.yiv6493947978msonormal41 {margin-right:0in;margin-left:0in;font-size:11.0pt;font-family:sans-serif;}#yiv6493947978 p.yiv6493947978msonormal311, #yiv6493947978 li.yiv6493947978msonormal311, #yiv6493947978 div.yiv6493947978msonormal311 {margin-right:0in;margin-left:0in;font-size:11.0pt;font-family:sans-serif;}#yiv6493947978 span.yiv6493947978EmailStyle34 {font-family:sans-serif;color:windowtext;}#yiv6493947978 .yiv6493947978MsoChpDefault {font-size:10.0pt;} _filtered {}#yiv6493947978 div.yiv6493947978WordSection1 {}#yiv6493947978
I don’t know if they made changes to the self test suite, but this unit goes into an extended diagnostics mode that very definitely tells you what FRUs (field replacement units…fancy euphemism for “we don’t do component level troubleshooting anymore”) may be responsible for each error code. I highly doubt those would be completely wrong, as that would waste the time of the service center people back when this unit was supported. There are *no* errors reported for executive functions or display functions.

 

However, that’s neither here nor there. I’ve now reseated everything, but no dice on clearing any errors.

 

Sean

 

From: Reginald Beardsley <pulaskite@yahoo.com>
Date: Saturday, March 27, 2021 at 12:39
To: Jerry Hancock <jerry@hanler.com>, Sean Turner <sdturne@q.com>
Subject: Re: error codes

 

You're assuming that the error reporting is correct. I'd advise not making that assumption.

There is not enough data to analyze the messages except in a very general way and we know almost nothing about what signals are on what connections.

KISS Pull each board and cable, then replace it. Go to the next board and repeat. Boot system. If errors show up repeat the process. That worked for my 11801 and Jerry's 11801C. It also worked for Jerry for the 11801 I got from him which is now throwing errors. I'll bet that when I do it tomorrow it will be happy again.

My bench is not completely clean, but very close.

BTW the 11801 loses the traces when powered off.



On Saturday, March 27, 2021, 01:25:08 PM CDT, Sean Turner <sdturne@q.com> wrote:

 

 

Yes, I see that now. However, there are no faults in those areas. All the errors point towards the acquisition module and possibly the strobe driver. I did figure out the acquisition module removal. I re-seated everything in the acquisition module; unfortunately this didn’t clear any of those faults.

 

Funny enough, the prior owner never cleared their data, so there are a bunch of stored traces available.

 

Sean

 

From: Reginald Beardsley <pulaskite@yahoo.com>
Date: Saturday, March 27, 2021 at 12:09
To: Jerry Hancock <jerry@hanler.com>, Sean Turner <sdturne@q.com>
Subject: Re: error codes

 


Yes, there are boards to pull. Look at p 6-14. The A17, A18, A14 & A15 all plug into the A13 motherboard. This is clearly documented on pp 6-50 to 6-56.

The 11801 has 2 memory boards instead of just one.

On Saturday, March 27, 2021, 12:47:48 PM CDT, Sean Turner <sdturne@q.com> wrote:

 

 

So there’s really no cards to pull, per se. Everything is connected with ribbon cables. My hypothesis is the problem lies in the acquisition module based on the self test results. I am working out how to remove it, as the manual isn’t perfectly clear on the matter. It (https://w140.com/tekwiki/images/d/de/070-8780-01.pdf) speaks of removing two Torx screws that hold a frame section in, but I can’t find any diagrams that show where these screws are.

 

Tomorrow is fine by me.

 

Sean

 

From: Reginald Beardsley <pulaskite@yahoo.com>
Date: Saturday, March 27, 2021 at 10:13
To: Jerry Hancock <jerry@hanler.com>, Sean Turner <sdturne@q.com>
Subject: Re: error codes

 


You don't need any sampling heads installed. I'm doing a CS class from 2-3 pm EST so tomorrow would be the soonest I could do it. I think Jerry is busy today also.

Did you pull all the cards from the backplane? IIRC they need to be in a specific order which is shown in the manual. It's quite a chore to reseat everything.

IIRC I was not willing to dig as far as the head connector board because of the huge PITA it presented.

I am quite amazed at how broken the autotools configuration for gnuplot-5.4.1 is trying to build on Ubuntu 14.04 LTS. God what a mess. It's depending on pkg-config which is broken. I think that imake has risen from the dead with a new name.

On Saturday, March 27, 2021, 10:55:55 AM CDT, Sean Turner <sdturne@q.com> wrote:

 

 

Yeah, that sounds good. Let’s do that once I get that sampling head you’re sending me, so I can rule out the thing freaking out because there’s no head present.

 

I’ve reseated all the connectors I can get to (seems to be mostly ribbon cables so far), but I haven’t yet figured out how to get to into the plug in bay…I will work that problem today.

 

Sean

 

From: Reginald Beardsley <pulaskite@yahoo.com>
Date: Saturday, March 27, 2021 at 08:32
To: Sean Turner <sdturne@q.com>, Jerry Hancock <jerry@hanler.com>
Subject: Re: error codes

 


The T3314 is most likely bad connection. At least that is what Jerry Hancock found recently with his 11801C. I had similar experience. I had to completely reseat my 11801 twice to clear the errors. I should note the errors were not shown in the photo on ebay. Just the NVRAM error. These beasts do not like traveling. I think they are probably all bad connections.

I'd suggest Deoxit, but getting at the backplane would be such an ordeal that just repeated reseating seems more reasonable. My first 11801 has been completely reliable once I had it working.

I'd like to suggest that the 3 of us get on the phone together to discuss. I suspect that among us we can deduce the answers.

I can drag my spare 11801 in once I clean off my bench and work on getting it going. It worked when Jerry foolishly sold it to me, but did not when it finally got here.

On Friday, March 26, 2021, 10:00:07 PM CDT, Sean Turner <sdturne@q.com> wrote:

 

 

These are the codes I pulled out of the selftest:

 

|
Subsystem
|
Block
|
Area
|
Routine
|
Index
|
|
Time Base
|
M/F I/F
|
Strobe Gate
|
Acq 1
|
T3314
|
|
Main Acq
|
Acq 1
|
Signal Path
|
Strobes
|
m1422
|
|
Main Acq
|
Acq 1
|
Signal Path
|
Signal Path A
|
m1461
|
|
Main Acq
|
Acq 1
|
Measurement
|
Filter A
|
m1521
|
|
Main Acq
|
Acq 1
|
Measurement
|
A->B B->A
|
m1541
|


Reginald Beardsley
 

In an attempt to get access to the calibrator I removed the nuts holding it in place per the service manual instructions. In the process I made a major tactical error and dropped one of the nuts into the instrument!

If you need to remove the calibrator for any reason, loosen the nuts and then use a multi-prong grabber to remove them. It's also advisable to have the instrument resting with the RHS on the bench. That will prevent the nuts from falling into the instrument.

As a consequence I now have both covers off the instrument at the same time. Fortunately, the nut fell out when I rotated the instrument.

The calibrator pin header is connected to J91 on Timebase Controller which is much more accessible for probing.

brown ???
red +4.93 V
orange ???
yellow ground
green ???
blue -4.93 V
purple ???

I have a clean square wave on the 3.5 mm line going to the calibrator. Does anyone know what signals should be on the brown, orange, green and purple wires? All I see with a scope is mV level noise.

Access to the signal path is through the RHS and the top. Both covers have to be off. The far RHS board is the TDR/Strobe buffer. The sampling heads plug into an interconnect board via ribbon cables. It's unclear from the manual if you can reseat the connectors in the signal path from the RHS after removing the TDR/Strobe buffer. But that looks to be the case. As I don't have signal path issues I don't want to mess with it.

I pulled the I/O board. As I expected the self test was turned off. In fact, all the jumpers were set to zero. I set them all to 1 in hopes of getting 19,200 baud on the RS-232 port, but it still reports 9600. Does anyone know if they will actually do 19,200?

I'm now throwing a T1331 error. That should be cleared if it boots normally. I saw this with #1 after I replaced the NVRAM. Do I need to jumper this to not halt on an error and then rejumper it? Pulling the card from the cage is a PITA. I remember the error from when I replaced the NVRAM in #1, but it cleared after a couple of boots. In this case, that is not happening. #1 is on a shelf 4+ ft from the floor, so I really don't want to wrestle it down to take apart to compare stuff.

Reg


Reginald Beardsley
 

FWIW In the event that you have a system on which the calibrator output is not working, if you connect the "Internal clock output" to an input and set the trigger to "internal" you should get a proper square wave, though at ~1 ns rise time in my case. Spec is <2 ns in the service manual. You can also verify the clock signal to the calibrator by connecting the 3.5 mm series coax to a sampling head. It should look just like the "internal clock" front panel output.

I'd like to note that I count 4-6 push on connections from the sampling head to the card cage boards. so there is lots of room for a fault which can be cured by reseating the connectors. Though a bit of PITA to get at some of them.

NB: I'm using an SD-22 for working on this. Nominally that's a 30 ps rise time head.

Reg