Racal Dana 9008 Modulation Meter woes.

[kkyahoo]

When cold my 9008 Racal Dana mod meter it does not lock onto a 10 Meg signal at 50mV. Lock light flashes. When warmed it is ok. Also there is a big 2.5kHz indication when cold too with no mod. Both these go when the unit has warmed up. Any ideas?

Thanks

Kevin

[Station X]

I can supply you with the service manual if you don't have it.

[G0HZU_JMR]

My old Racal 9008M became unreliable a few years ago and it had vaguely similar symptoms. I noticed significant problems with the AM accuracy when cold. It had to be on for >30minutes to begin to give sensible readings.

When I took it apart, it was easy to spot that nearly all of the 47uF 16V electrolytic caps on the lock assembly board had leaked. Each one had a faint 'watermark' stain around it on the PCB. When I removed them the ESR of these caps was >20 ohms. Some were much higher than this.

I think these were caps C10, 14, 23, 33, 36, 39, 43, 57 and C58. I tried to buy low ESR replacements and went for Panasonic 'gold' types but the ESR of the new caps was a bit disappointing. It was something like 1 or 2 ohms. However, I still fitted the new caps and the meter now works fine again, even when cold. I keep meaning to find some 47uF 16V caps with better ESR because I think this wonderful little meter deserves this.

[G0HZU_JMR]

If you do find that these caps are leaking on your lock assembly then I would strongly recommend you take plenty of photos of the board before you remove the old ones. The assembly drawing in the manual doesn't easily show which way round they should be fitted. So an easy job could turn into a major headache if you have to try and reverse engineer which way round the caps should be fitted. I think the PCB has some + marks on it to show the polarity but I recall that these aren't placed very well on the PCB.

[AC/HL]

1-2 ohms is reasonable for this value.

[G0HZU_JMR]

I know but I think the Panasonic datasheet for the caps I chose promised an ESR of <=0.8R at 100kHz. I chose these caps quite carefully. So I was a bit disappointed when I measured them. My ESR measuring setup is normally very good and I'd expect it to measure this ESR quite accurately. It can measure 47uF tant caps with much lower ESR for example.

Maybe the ESR of the Panasonic 47uF 16V electrolytic caps improves once they see a bit of real use?

I guess none of this matters if this isn't relevant to kkyahoo's meter. However, because nearly all of these caps had failed in my meter there is a chance this is a common problem. I imagine quite a few of these meters will have had these caps replaced already. So any visible PCB stains around these caps could be from a previous set of caps.

[kkyahoo]

Thanks for all your replies...I too have inaccurate a.m readings. I will have to get the lid off and look at the caps, I am guessing it will be a cap somewhere!
I looked at the IF output round the back and it is very noisy until warm. I think the psu is ok as connecting the 9008 to the mains and leaving it switched off makes no difference to the fault.. in this condition the mains psu is powered up but nothing else is. The measuring board looks ok as it is doing what you expect, that kind of implies the lock board is dodgy.

Kevin

[kkyahoo]

I looked at the lock board and it was showing signs of being attacked by leaking capacitors. The caps were measuring 5 to 20 uF, ESR in the range 5 to 20 ohms. The replacements measured 1 ohm ESR and the capacitance was fine at approx 47 uF.

9 capacitors later and with the unit reassembled it was fully working again! The AM in particular was exactly right after reading high before it decided to fail.

Thanks for all the replies especially from Jeremy and thanks Graham for the service manual!

The 9008 is now fully working.

Kevin

[G0HZU_JMR]

Glad it's fixed! I too had the symptom of the AM reading high when cold. It would gradually drift down to the correct value after maybe 30 minutes or so. Once I changed all the 47uF 16V electrolytic caps the AM reading went back to being spot on even when cold. So these symptoms are probably going to be quite common as these meters get older and older.

[kkyahoo]

Thanks Jeremy,

It's a great piece of kit. Very clean it seems too. With no mod on input just 100mv carrier it's residual mod is 0.025% on AM, 0.025kHz on FM, are these what you are seeing?

Kevin

[G0HZU_JMR]

In my case it depends on which sig gen I use. The Marconi 2024 has very low carrier noise and low spurious down at a few MHz so with this sig gen with a clean carrier at 100mV the 9008M needle hardly moves even on the 1.5 range on AM or FM. I get about 17mV at the DC/DMM port at the back but I'm not sure if this needs to be zeroed before I do the test.

I think one of the best features of the meter is the quality of the internal AM and FM demodulators. They both give remarkably low distortion at the AF output port at the rear of the meter. I'm not sure if my sig gens are the limit or the 9008 but I typically see distortion terms >60dB down and this is quite impressive.

[kkyahoo]

Thanks Jeremy,
Looks like there are 2 more electrolytics in the sampler, C15 and C16, are they worth changing too? Would they effect the reading? I am seeing a little drift on FM only, 2.9 k dev'n drifts to 3k with a 3k input. Maybe not significant.

Kevin

[G0HZU_JMR]

On my meter those electrolytic caps look very different in terms of packaging (different manuf/series?) and they appear to test out OK for ESR. So I have left the originals in place. But it is definitely worth checking these caps on your meter just in case they have failed.

[G0HZU_JMR]

I've just done a fairly harsh test on my 9008M meter to see how much it drifts from cold. Today is a really cold day and the meter was stored in a back room. So the meter was quite cold when I picked it up. I tried testing it from a cold start on 50% AM at 11.8MHz.

I have two ways of making sure the mod depth of the test sig gen is accurate. I measure the relative depth of the mod sidebands to the carrier with an analyser that has a digital IF. This has a very accurate log scale. I also have a real time spectrum analyser that has an exotic modulation analysis option fitted. So it can measure AM mod depth (peak, trough, pk-pk etc) with a meter that operates in the digital domain. Both setup methods agree with each other within 0.1% of 50%. I also have a steep LPF after the sig gen to remove any harmonics or high frequency spurious signals from the sig gen.

When I first switched on the 9008M it read 51.7% on a DVM via the DC recorder at the back. Over the following 45 minutes it gradually drifted down to read 50% +/- 0.2%. Not bad for an old meter and I think it is fair to allow it 30 minutes to stabilise. For the last 10-15 minutes it has stabilised at 49.8% on the DVM.

When my meter was poorly with the duff 47uF caps it would read well over 90% mod depth for an accurate 80% test signal at first switch on. It would then gradually drift down over time to read closer to 80%. So it was definitely not working well at all.

[kkyahoo]

Thanks Jeremy,

I noticed that the residual mod changes with the af filter switched out of circuit if I switch the power on and off again. Mine's a 9008M version too.
I will report back if I change anything else.

Kevin

[G0HZU_JMR]

OK but just for completeness sake I repeated the above tests for FM. I left the 9008M meter unpowered in a cold place for over 2 hours and then retrieved it and powered it up and tested it at an accurate 2.5kHz FM deviation. This deviation was set with a Bessel null and the AF source was synthesised from a highly accurate OCXO. The FM deviation was then checked on the accurate demodulator on the RTSA to be 2.500kHz +/- 0.005Hz.

The difference between a cold start and after a decent warmup time was fairly negligible. On the DC output from the 9008M the DVM read 0.494V when cold and this corresponds to an indicated 2.47kHz deviation.
When warmed up it changed to 0.498V. This indicates 2.49kHz FM deviation. After another 45 minutes' warmup it had hardly drifted at all. So I think my meter is very stable in FM mode. It's not so good in AM mode but probably still in spec. It's definitely in spec if a 30 minute warmup time is allowed.

However, I generally use this meter as an accurate AM or FM detector/demodulator via the AF output connector at the rear of the meter and I rarely use it to measure FM deviation or AM depth on the front dial. I really do like these meters and they are beautifully built inside with very neat wiring etc.

[kkyahoo]

Thanks Jeremy,

Great input. You have some nice gear there!

I suspect on my 9008 that the sampler and PSU electro caps need changing but that is just a hunch!

Kevin

[kkyahoo]

I ordered some further new caps and checked the residual noise again. With the af filter out of circuit using FM the residual noise was 200Hz going down to half this if the mains power is blipped. With AF filter engaged it just met spec of 25Hz. (Generator is 10 times better than this, 100mV input used). I changed the frequency to 10MHz then to 145MHz and the noise went up to 200Hz again and blipping the mains on and off made it halve once more, so the effect is repeatable. Don't know if they all do this? Does it on AM to albeit different values.

Kevin

[G0HZU_JMR]

200Hz sounds quite high assuming you are on the 1.5kHz range on the meter. If it reduces to 25Hz with the AF filter(s) switched in then it might be worth having a look at the signal at the AF output BNC at the rear of the meter when you are doing this test. This port should be compatible with a PC soundcard input so if you can download some freebie spectrum analyser SW you could have a look at the demodulator output to see if it is clean from a few Hz to maybe 20kHz. The other option is to use a scope to look here especially if it is a DSO with an FFT function built in. I'm not sure which AF filter you are using (150Hz or 3kHz) but I suppose it could be filtering out high frequency noise/spurious or maybe there could be 50/100Hz here also.

[Biggles]

Reading this thread has encouraged me to dig out my 9008 and get it going again. I haven't used it for a while now, but remember that this was a common machine in the eighties, before the likes of the all-in-one test sets came along. We used to connect all the seperate test gear up on site then give it half an hour (plus) to warm up and settle down. That was the excuse to go and have a cuppa and read the paper anyway. I remember it was a bit of a job connecting everything up, with each piece of gear contained in it's own protective custom made box. So much easier nowadays. A lot of the time you just seem to need a laptop and not much else.
Alan.