|
Vintage Test Gear and Workshop Equipment For discussions about vintage test gear and workshop equipment such as coil winders. |
|
Thread Tools |
29th Apr 2008, 9:43 am | #1 |
Dekatron
Join Date: Aug 2003
Location: near Reading (and sometimes Torquay)
Posts: 3,100
|
Marconi TF2330 AFC problem
I am running out of time to fix a problem with the crystal discriminator used to do the AFC on this old wave analyser, and I'm getting desperate to solve this remaining problem.
Having fixed a more serious fault caused by a duff capacitor, I thought it was working but not as well as I would have hoped. The obvious symptom is that while it locks very well from one side, it is very very slow the other way although it gets there in the end, in spite of the voltages from the discriminator itself being very plausible. To make it easier to see what voltages were being generated I tried shorting out the output offset voltage, and this revealed that merely having the AFC circuit active was deviating the local oscillator by 10mV of control voltage, which means that the discriminator must counter this so the zero frequency deviation point is not at zero offset voltage. Surely this isn't right? I can't see where this offset comes from. The trouble is that the offsets are so small, it's <10mV out of the standing 13.13V control voltage but that makes a good 3% reading error, and connecting test equipment is usually enough to mess it up completely. |
29th Apr 2008, 10:06 pm | #2 |
Retired Dormant Member
Join Date: Oct 2004
Location: Bristol, UK.
Posts: 837
|
Re: Marconi TF2330 AFC problem
I've no experience of this piece of kit but since you are running out of time:
1): Look for any signs of tinkering in the past, changed components, altered wiring and so on. Put a scope on the supply lines and decoupled supply lines and make sure there's no motorboating. 2): Dry joints can cause odd problems; remake the joints, especially earthing. 3): Just swap all the parts that are easily changeable and which might affect this. Old caps can have really odd behaviour. 4): Anything on the WWW? 5): This works with software problems. Person-to-person, explain the problem to someone else, preferably someone who has knowledge of electronics, next choice someone who solves problems and asks questions, last choice, anyone who will sit down and listen. With software problems very often you spot the issue as you explain it. I think it's something to do with being stuck in a rut of thought and the process of having to explain it makes you look at it from a fresh perspective. Explaining it via text on the forum isn't the same. 6): You're not possibly trying to make this work better than its original spec? 7): Read the service manual again, this time like a love letter, looking even for the merest hint. HTH Pete. |
29th Apr 2008, 11:04 pm | #3 |
Dekatron
Join Date: Aug 2003
Location: near Reading (and sometimes Torquay)
Posts: 3,100
|
Re: Marconi TF2330 AFC problem
Yes, it may well be that it is working as well as it does, but I am having difficulty reconciling its slightly odd performance with measurements of its workings.
One wretched problem is that I can't replace the final capacitors that smooth the AFC because I can't find modern replacements that are anywhere near good enough. It uses a pair of 330uF 6V electrolytics back to back - but these are clearly special low leakage ones and indeed they have very low leakage (0.5uA at 5V). The problem is that they may be playing up at the voltages being used, i.e. +-10mV. I have tried charging them up and this does seem to make a slight improvement, but nothing dramatic. My best guess so far is that there is some asymmetry to the input waveform and this results in a tiny shift in the average power consumption which drops a few mV across the decoupling. It may be this changes as the crystal goes from the series to parallel resonance points and is thus cancelling part of the control voltage. The trouble is these voltage changes are less than 0.1% which is at the limit of what I can measure. |
29th Apr 2008, 11:52 pm | #4 |
Retired Dormant Member
Join Date: Oct 2004
Location: Bristol, UK.
Posts: 837
|
Re: Marconi TF2330 AFC problem
60's transistor radios and other kit often had 6V or 3V electrolytics which you don't often see now, but check the catalogues.
If they are truly special electrolytics, now out of production, you might just have had it. What about replacing them with tants? Can't help much with regard to the assymetry of the input waveform, but........ Why do you think it's assymetrical? Scope trace, hunch, necessary conclusion? What's causing the assymetry? Can you find a way of inverting the input and if so is the locking behaviour is reversed; slow from the other side? If it's inverted and things stay the same, it isn't assymetry causing this. Why exactly should leakage in the caps cause this? Are they both equally leaky? What happens if you tinker with the leakage, say by clipping a 1M resistor across the caps? What happens if you tinker with the leakage to make it the same in both cases? Pete. |
30th Apr 2008, 9:53 am | #5 | ||||
Dekatron
Join Date: Aug 2003
Location: near Reading (and sometimes Torquay)
Posts: 3,100
|
Re: Marconi TF2330 AFC problem
Quote:
Quote:
Quote:
Quote:
|
||||
30th Apr 2008, 10:54 am | #6 |
Retired Dormant Member
Join Date: Oct 2004
Location: Bristol, UK.
Posts: 837
|
Re: Marconi TF2330 AFC problem
Nichicon do very low leakage caps (0.4µA), which are available from Mouser in the US.
http://products.nichicon.co.jp/en/pdf/XJA043/e-ml.pdf http://www.mouser.com/Search/Refine....user_Wildcards Maybe you can get them from Farnell or RS, if you search there with the Nichicon part number. Possibly other leading makers of electrolytics do similar or better. Pete. |
1st May 2008, 9:23 am | #7 |
Retired Dormant Member
Join Date: Mar 2008
Location: Hampshire, UK.
Posts: 297
|
Re: Marconi TF2330 AFC problem
Have just measured some of the small 150µF 6V "blob" tantala. Out of 12 or so only one was high at 15µA the remainder were in the 0.2 to 0.4µA range. Tested at 5V after 30 seconds.
Just try - alternatively "prove" by finding something to degrade in the same way? Adding leakage by parallel R - does that have "same" effect? Roy |
1st May 2008, 11:03 pm | #8 |
Dekatron
Join Date: Aug 2003
Location: near Reading (and sometimes Torquay)
Posts: 3,100
|
Re: Marconi TF2330 AFC problem
That result is what I would have hoped for, but such high value tantalums seem both hard to find and very expensive.
In desperation I have indeed fitted a 5M6 resistor to cancel the leakage current that must be flowing. I looks like it may be the varicap itself. It's a 100SC2 but I can only find US suppliers and expensive too. With the leakage compensation in place (represents about 0.1-0.2uA) results are not too bad with the remaining issues being down to either the back to back capacitors or maybe the crystal filter characteristic is slightly off. Anyway it is probably as good as it is going to get now, my only big concern is that if the leakage changes it will go out of alignment again. I have the offset down to 40mV now. One concern is that this leakage effect clearly changes quite a bit during warm-up so I will need to leave it on for a long while and with the covers on to check this will be OK. This is not the first time I have worked on a TF2330 and I think it is the hardest thing to work on that I have ever encountered. It pushes my test equipment beyond the limits, where 0.1Hz or 1 mV makes a massive difference! |