Author Topic: Racal Frequency Counter  (Read 11599 times)

0 Members and 1 Guest are viewing this topic.

Offline veedub565Topic starter

  • Frequent Contributor
  • **
  • Posts: 407
  • Country: gb
Racal Frequency Counter
« on: July 15, 2015, 07:37:50 pm »
Hi I'm trying to fix an old Racal 1992 Frequency counter (option 04C) and was hoping someone might have some more experience on this than me. Here's how far I have got so far.

Racal-Dana 1992 (option 04C) - Does not boot up, when you switch it on it displays random lights, characters, sometime everything lights up sometimes nothing lights up. All power supply rails and clocks are ok, 10mhz reference is coming out of the back ok.

It is supposed to display model number and software version when it boots up, but since it wasn't even getting this far I suspected the EPROM. I was able to swap the EPROM from another identical known good unit and my faulty one now boots as far as displaying the model number (1992) but then it freezes there and will not go any further.

It's obviously a issue with the EPROM causing the original fault, but I cannot now understand the new fault. Both EPROMs have exactly the same software version sticker on them so they ought to be identical? If it works in one it ought to work in the other.

Any idea's ?
 

Online Andy Watson

  • Super Contributor
  • ***
  • Posts: 2074
Re: Racal Frequency Counter
« Reply #1 on: July 15, 2015, 08:02:50 pm »
Ha ha! So you're the one that out bid me?
Clearly the Eprom is ok. Remember that Eprom will be sharing the bus with the display/keyboard chip and custom timer chip - if the selection logic for either of these is broken they could be fighting for control of the data bus. Given the half-filled digits shown in the photo I would be concentrating on the display/front panel - at least if that works it would give it half a chance of displaying a fault message.
« Last Edit: July 15, 2015, 08:04:23 pm by Andy Watson »
 

Offline veedub565Topic starter

  • Frequent Contributor
  • **
  • Posts: 407
  • Country: gb
Re: Racal Frequency Counter
« Reply #2 on: July 15, 2015, 08:10:57 pm »
lol I wondered if anybody would recognise it, for that price it was worth a punt. Looks like it may be a paperweight though.

The (original) EPROM must be duff as it won't even post the model number at boot And when I put the EPROM in from another unit then it gets a bit further along the boot process and does display the model number. The display works fine.

 

Online Andy Watson

  • Super Contributor
  • ***
  • Posts: 2074
Re: Racal Frequency Counter
« Reply #3 on: July 15, 2015, 08:22:57 pm »
The difference in Eproms might just be down to their ability to drive the bus and/or slightly quicker reaction time to chip select etc.. The seller said something about the eprom chip select being dodgy - I think I would look further afield at everything that shares the data bus. I assume you've got the service manual: http://www.ko4bb.com/manuals/index.php?dir=Racal
 

Offline veedub565Topic starter

  • Frequent Contributor
  • **
  • Posts: 407
  • Country: gb
Re: Racal Frequency Counter
« Reply #4 on: July 15, 2015, 08:39:42 pm »
Yeah thats a possibility I hadn't considered, and as you say if multiple chips are fighting for control of the bus that would cause it to freeze up. I checked out the EPROM chip output enable (active low) as this is what the seller was on about. It originates from the MPU and with the original EPROM it's high, with the new EPROM it's low.

I also swapped over the MPU and custom ASICS from the working unit just to be sure the originals are ok. I'm going to check the original EPROM in the reader/writer just to see if it can be read ok (or not)

I have a do have a paper copy of the service manual.

 

Offline veedub565Topic starter

  • Frequent Contributor
  • **
  • Posts: 407
  • Country: gb
Re: Racal Frequency Counter
« Reply #5 on: July 15, 2015, 09:37:55 pm »
Clutching at straws here, but, one small difference between the faulty unit and the known good unit is the faulty on has the 04C option (basic crumby 10mhz) and the known good one has 04T option (better tcxo with oven?) Maybe, although the eproms look identical and have the same software ver. Maybe they are still programmed slightly differently, maybe when it boots up it checks the accuracy and/or stability of the internal 10mhz ref. The eprom from the working unit will be expecting to find the oven tcxo quality ref, and when it doesnt it sits and waits for it to settle down before allowing measurements to take place. Hence apparently freezing/not booting.

Will try swapping over the oven tcxo tmoz and let you know. Long shot but you never know.
 

Offline grumpydoc

  • Super Contributor
  • ***
  • Posts: 2905
  • Country: gb
Re: Racal Frequency Counter
« Reply #6 on: July 15, 2015, 09:43:09 pm »
Quote
Clutching at straws here, but, one small difference between the faulty unit and the known good unit is the faulty on has the 04C option (basic crumby 10mhz) and the known good one has 04T option (better tcxo with oven?)
04T is just a TCXO, no oven

04E is the high stability oven - you want this if you can get it

04A/B are the standard OCXO options, the A being fast warm up and B being slightly more stable.

04C is a basic crystal, intended for use with an external standard

Quote
Maybe, although the eproms look identical and have the same software ver. Maybe they are still programmed slightly differently, maybe when it boots up it checks the accuracy and/or stability of the internal 10mhz ref. The eprom from the working unit will be expecting to find the oven tcxo quality ref, and when it doesnt it sits and waits for it to settle down before allowing measurements to take place. Hence apparently freezing/not booting.

Will try swapping over the oven tcxo tmoz and let you know. Long shot but you never know.
Unlikely - the oscillator in these just sits on a separate board at the back with a connector that carries power and the 10MHz back - there's no way for the CPU to know what type it is.

In fact I'm pretty certain the firmware is the same for all the 199x counters - it works out what model from which front panel is present

Oh, it's worth replacing the PSU caps in these models, they are a known weakness especially the big 680uF ones.

I know you said power rails were OK, but double check them for ripple.
« Last Edit: July 15, 2015, 09:49:18 pm by grumpydoc »
 

Online edavid

  • Super Contributor
  • ***
  • Posts: 3381
  • Country: us
Re: Racal Frequency Counter
« Reply #7 on: July 15, 2015, 09:50:10 pm »
I know these units are prone to keypad problems... does a stuck switch cause them to hang?
 

Offline grumpydoc

  • Super Contributor
  • ***
  • Posts: 2905
  • Country: gb
Re: Racal Frequency Counter
« Reply #8 on: July 15, 2015, 09:57:03 pm »
I know these units are prone to keypad problems... does a stuck switch cause them to hang?
I just tried powering my bench 1992 up with a key pressed - it will stop any other keys working but the unit still boots
 

Online Andy Watson

  • Super Contributor
  • ***
  • Posts: 2074
Re: Racal Frequency Counter
« Reply #9 on: July 15, 2015, 10:07:08 pm »
In fact I'm pretty certain the firmware is the same for all the 199x counters - it works out what model from which front panel is present
Yes, I think that's right. The 1991/1992 software is identical - it detects the difference by link on the HF module. The clocks are interchangeable on plug-in modules - as is the IEEE/GPIB and battery option - it's all the same firmware.
 

Online edpalmer42

  • Super Contributor
  • ***
  • Posts: 2260
  • Country: ca
Re: Racal Frequency Counter
« Reply #10 on: July 15, 2015, 10:18:52 pm »
One thing I've found with my 1992 is that there must be an internal clock - even if you've got it connected to an external standard.  On bootup, the internal clock is used until it recognizes that an external reference is connected.  Then it switches over to the external standard.  So, if the internal clock is flakey, the unit might not boot properly.

 

Offline lowimpedance

  • Super Contributor
  • ***
  • Posts: 1245
  • Country: au
  • Watts in an ohm?
Re: Racal Frequency Counter
« Reply #11 on: July 16, 2015, 03:32:32 am »
One thing I've found with my 1992 is that there must be an internal clock - even if you've got it connected to an external standard.  On bootup, the internal clock is used until it recognizes that an external reference is connected.  Then it switches over to the external standard.  So, if the internal clock is flakey, the unit might not boot properly.
I have a couple 1991s and one which does not get used all that often does hang on boot up initially atfer sitting for some time unpowered. This behaviour is generally repeatable too after some months of no power !!.
 Suspected a dodgy connection on the timer IC socket since this beasti gets quite hot in normal operation so thermal cycling issues and given its age !. But now you mention the clock perhaps thats also suspect at the connector !.
 Usually the unit starts working once the cover comes off so tracing the culprit has not been definative. (and yes I have changed the socket, not that!)

And definately change the 680uF caps.
The odd multimeter or 2 or 3 or 4...or........can't remember !.
 

Online edpalmer42

  • Super Contributor
  • ***
  • Posts: 2260
  • Country: ca
Re: Racal Frequency Counter
« Reply #12 on: July 16, 2015, 05:07:49 am »
I have a couple 1991s and one which does not get used all that often does hang on boot up initially atfer sitting for some time unpowered. This behaviour is generally repeatable too after some months of no power !!.
 Suspected a dodgy connection on the timer IC socket since this beasti gets quite hot in normal operation so thermal cycling issues and given its age !. But now you mention the clock perhaps thats also suspect at the connector !.
 Usually the unit starts working once the cover comes off so tracing the culprit has not been definative. (and yes I have changed the socket, not that!)

And definately change the 680uF caps.

If by 'timer IC' you mean IC39, a ceramic IC that normally runs smoking hot, that's the chip that detects the external reference and switches over.  They're custom chips that are used in all the Racal 199x counters.  I've got two units that are dead because that chip died.  I've thought about replacing it with a CPLD or FPGA, but since I can barely spell those names, it's probably not going to happen.   :)

 

Offline lowimpedance

  • Super Contributor
  • ***
  • Posts: 1245
  • Country: au
  • Watts in an ohm?
Re: Racal Frequency Counter
« Reply #13 on: July 16, 2015, 06:54:48 am »
Heh yeah thats it, I was too lazy to check the manual for the ref./ desrcription.
On the unit that I do use more regular I made a sort of heat pipe type of contraption the bolts to the side of the chasis and is adjustable to apply light pressure contact to the top of the IC and hopefully conduct some of that HEAT away, (does seem to work!)
Perhaps I could post a picture if any one is interested .
The odd multimeter or 2 or 3 or 4...or........can't remember !.
 

Offline veedub565Topic starter

  • Frequent Contributor
  • **
  • Posts: 407
  • Country: gb
Re: Racal Frequency Counter
« Reply #14 on: July 16, 2015, 07:36:59 am »
Thanks for all the info guys. I have now confirmed for sure that the original EPROM is dead, I plugged it into an external reader and it could not read anything from it. So it looks like that's one fault dealt with.

I'm making a copy of the working EPROM to plug into this unit now.

So now the issue is why does it freeze during boot, and could this be what blew the EPROM in the first place (hence making a copy)

Plan now is to:
- double check the power lines for ripple
- check the internal 10mhz reference
- and also have a look at the chip select/address lines.
 

Offline veedub565Topic starter

  • Frequent Contributor
  • **
  • Posts: 407
  • Country: gb
Re: Racal Frequency Counter
« Reply #15 on: July 16, 2015, 09:11:01 am »
Ok an update:

- Power rails look ok and reasonably ripple free.
- 10Mhz internal ref looks ok

Looking at the address lines on a scope shows they are clocking high/low ok.

I also had a look at the data lines, and it looks like there is two different signals one on top of the other. I've ruled out (by way of replacement) IC22, IC18, IC19, and IC39 so that only really leaves IC20

I'm guessing it's possible IC20 is constantly sqwaking on the data line, talking over the top of other chips and not shutting up when it's supposed to.

I've removed IC20 and ordered a new one which should arrive in a day or two. Will let you know how it goes.

 
 

Offline grumpydoc

  • Super Contributor
  • ***
  • Posts: 2905
  • Country: gb
Re: Racal Frequency Counter
« Reply #16 on: July 16, 2015, 12:40:57 pm »
What 'scope do you have

Don't forget that if you look at a typical MPU data bus with a 'scope you tend to see what looks like a jumble of signals because the timing can be quite irregular.

the best way to look at a data bus is, if you have a DSO with decent memory depth, is to capture activity as a single shot then expand the time base and scroll back and forth.

If you have an analogue 'scope and the CPU is executing just a small bit of code in a loop (eg waiting for a key press) then you can sometimes get a stable trace by adjusting the hold-off control.

If multiple devices are trying to drive the bus at the same time and one is at a "1" with the other at "0" the bus will take up an intermediate value (and things will get warm).

However also remember that the bus only needs to be at a "clean" data level at a specific point in a read or write cycle and you get periods when nothing is actively driving the bus so it can take up odd values at those times as well.

 

Offline veedub565Topic starter

  • Frequent Contributor
  • **
  • Posts: 407
  • Country: gb
Re: Racal Frequency Counter
« Reply #17 on: July 16, 2015, 01:01:44 pm »
I'm using a 500MHz HP Infinium

I did check the working unit and the data lines looked pretty similar to my faulty unit so it may not be IC20, at least it will rule it out though once it's replaced. I've also ordered IC21 which is the chip select latch.  That then eliminates a lot of the microprocessor system.  Won't have the new parts till next week now.

It's a shame there isn't a documented boot sequence so at least we'd know what its doing when it boots.
 

Online Andy Watson

  • Super Contributor
  • ***
  • Posts: 2074
Re: Racal Frequency Counter
« Reply #18 on: July 16, 2015, 01:30:23 pm »
I don't think IC20 dirves the data bus (unless it's properly broke). It only has inputs on the data bus.  Have you checked that the chips have power and ground on the actual pins of the chip?
« Last Edit: July 16, 2015, 02:08:55 pm by Andy Watson »
 

Offline veedub565Topic starter

  • Frequent Contributor
  • **
  • Posts: 407
  • Country: gb
Re: Racal Frequency Counter
« Reply #19 on: July 16, 2015, 03:35:45 pm »
Checked the chips around the MPU circuit (IC39,18,19,20,21,22,23,24,25) and they all have 5v and gnd on the chip itself.

Won't be able to do much more now until the new parts are fitted next week as I've already removed IC20.
 

Offline gfiber

  • Contributor
  • Posts: 35
  • Country: us
Re: Racal Frequency Counter
« Reply #20 on: July 16, 2015, 09:25:02 pm »
You might take a look at Gerry Sweeney's blog www.gerrysweeney.com if you have not. He had a Racal 1999 counter that would lock up. It appeared to be a faulty break or connection on the main board but turned out to be a faulty 10 Mhz Osc connector. In another video he built his own design High Stability Osc for that unit and also replaced all the switches. Ian Johnston at www.ianjohnston.com replaced all the switches in his Racal 1992, but I do not remember him having any other issues than bad switches.

Some of the above may help
Gary K8IZ
 

Offline rastro

  • Frequent Contributor
  • **
  • Posts: 388
  • Country: 00
Re: Racal Frequency Counter
« Reply #21 on: July 17, 2015, 02:50:35 pm »
I purchase the same model with the same problem and it turned out to be a stuck key(s).  The key caps can be carefully pulled/removed from the front panel.  Then remove the rubber switch contact and ensure there is no debris between the gold contacts on the underling circuit board.  This should take less than 15 minutes.  I would do this before starting to trace signals.

-rastro
 

Offline veedub565Topic starter

  • Frequent Contributor
  • **
  • Posts: 407
  • Country: gb
Re: Racal Frequency Counter
« Reply #22 on: July 21, 2015, 10:34:20 am »
Well it isn't IC20

I tried swapping the entire front panel assy from another working unit and it didn't make a difference so that rules out a stuck key or display fault.

back to the drawing board.
 

Online Andy Watson

  • Super Contributor
  • ***
  • Posts: 2074
Re: Racal Frequency Counter
« Reply #23 on: July 21, 2015, 09:42:39 pm »
That doesn't leave much! Do you have a scope with enough channels to monitor the address decoding gates - they're only two-input, output ors?
 

Offline veedub565Topic starter

  • Frequent Contributor
  • **
  • Posts: 407
  • Country: gb
Re: Racal Frequency Counter
« Reply #24 on: July 22, 2015, 09:47:44 am »
Not to hand no, but I could borrow a 4 channel one.

I had another look at the supply lines again, initially dismissed the ripple as it was quite small (200mv p-p) but I had another look today and compared with the working unit. On my unit the ripple is 50hz and on the working unit it's 100hz. My theory's a bit rusty but 100hz is full wave rectification and 50hz is half wave, so maybe the bridge rec is faulty.  I had a look at it and D12 is 30k ohms both ways across pins 2-3 compared to 2Mohms on the working one. So may be worth having a closer look or replacing D12.

*Edit* Replaced D12 and it's not that. more investigation required.
 
« Last Edit: July 22, 2015, 12:53:44 pm by veedub565 »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf