Inductance measurement problems with a TF2700

[Skywave]

Recently acquired a Marconi TF2700. Many, many years ago I used one of these quite often. Downloaded the manual for a 'refresher' and read it. Ah, yes: it all 'came back'!
The 'recent acquisition': on initial test, yes, it worked . . . sort of. So, replaced all the electrolytics and various resistors on the detector amp. board and ditto the 1kHz oscillator. Result: now O.K. for R and C, but not for several inductors. Example: an 80 µH R.F. choke and various other inductors / coils failed to produce even a hint of a null. All of these coils are readily measurable of my vintage Marconi TF868B LCR bridge and ditto on my HP LCR meter (Which has an analogue meter for indication and uses a switchable two-frequency Wein bridge).
With this TF2700, I tried an external a.c. source. But at 10 kHz and even 20 kHz, no improvement.

Am I missing something fundamental here? Any help / guidance will be appreciated.

Al.

[Diabolical Artificer]

My TF2700 doesn't measure inductance well either, also it struggles with some caps. It could be my technique though on the whole I find the TF2700 difficult and time consuming to use. It's a bit like cracking a safe; tongue at the right angle, don't breath, very, very, very gently turn the knob......

Andy.

[Ed_Dinning]

Hi Gents, these bridges can show strange effects depending in L/R ratios and to a smaller extent C/(leak) ratios.
Usual trick is to use 2 hands and adjust both Balance and R.balance together. This may sometimes lead you to "chase the null" right down to zero. If so use a different bridge configuration (Q switch).
Make sure slide wire is clean.
Once you get used to using it it is a good piece of equipment and possibly one of the best for price/ performance.
It far outperforms the L or C ranges on DVM's that seldom give accurate readings.

Ed

[Sideband]

I seldom use mine for inductance measurements...not because of any apparent difficulty/uncertainty but because I've had no need to.... Interesting...next time I'm in the workshop, I'll find some known inductors and check them.

[Skywave]

Gentlemen - thank you for your responses.
At present - and subject to further testing, etc. - I am not impressed with this item. In terms of performance and in-built facilities, it seems a poor quality derivative (if that's what is is) of the earlier, valve-based, TF868B.

Al.

[Skywave]

Originally Posted by Ed_Dinning

Usual trick is to use 2 hands and adjust both Balance and R.balance together. This may sometimes lead you to "chase the null" right down to zero. If so use a different bridge configuration (Q switch).

Thanks: I tried all the available 'user controls' with many coils / inductors that would not respond on this TF2700. It has the facility to use an external oscillator, so I tried that with several coils that had a relatively high R to L ratio, hence a low Q. The thinking there was that since inductive reactance is proportional to frequency and that a coil's R.F. resistance is also so dependent, but far less so, an exciting source freq. of 10 kHz to even 50 kHz should have given some sort indication, (because then XL will then be proportionally increased compared to its R, so its Q will be increased), even if that null is a shallow one - but no - no hint of any balance at all.

Al.

[G4_Pete]

HI,
One way I find useful to find a starting null on inductors using the TF 2700 is to approach the null down from a very high value as in the worked example on one of my Grid dip meter coils below.

1. set the range multiplier way above the expected value and the Q on X1.
2. Set the outer Click selector to Zero.
3. Set the Loss to about 1
4. Then on the L range you should find a Null around the 00 setting on the fine setting.Crucially this will be fairly insensitive to the Loss balance at this point which should be set around 1.

I have my TF2700 in front of me now with a 39 uH coil off of my dip meter on test.
With the range at 1H and loss at 1 there is a deep null at 00. Loss does little EXCEPT if you take it below Zero .

5. Dropping the range to 100 mH starts to show a null above the zero ( 00) point on the fine setting. The loss setting should now start to show a broad null. Mine is now showing a null at 0.5 on the loss and just above the zero on the fine control.
6. Drop the range to 10 mH . The null creeps up to fine scale and the loss balance becomes more critical . Persivere to find a good null at about half sensitivity .
7. Drop the range to 1 mH and WITHOUT touching the loss use the fine range dial to re find the null . I now get a null at 0 on the outer click range and 30 on the fine dial.
8. Having re found the null on the fine dial at the 1 mH setting the loss can be re adjusted to get a deep null but settings are getting critical now.

9. I now get a usable reading of 0 on the outer click setting and approx 39 on the inner fine setting.
10 Dropping once more now to the 100 uH range the outer ring can be set to 3 and the inner balances to 92 . Some fine control off the loss is possible but at this point mine is reading 0 on the loss setting. The loss is VERY critical at these settings.

Therefore the final value is .392 x 100 uH = 39.2 uH .

I hope this helps but I will concede these are very fiddly bridges to operate.

Pete.

[Radio Wrangler]

Boonton/HP had a competing small bridge using a 10 turn pot and a cyclometer counter as the scale. It's special feature was a phase detector looking at the unbalanced error voltage, operating two neons to indicate which way to turn the knob to find the null. It makes things a lot easier.

THese were all essentially instruments for component checking on the bench the older TF868 was more of a lab machine.

THE bridge to have is one of the Wayne-Kerr Autobalance jobs... the ones with two meters and banks of decade buttons for balancing. You select a range that gives you meter readings with all zeroes pushed, then the meters tell you which top buttons to press in the reactance and resistance banks, then the meters tell you the buttons for the next rows, then the third rows. You can either read the fourth digit off the meters, or back them off with calibrated 10 turn 'knobpots'

Brilliant machines!

David

[G4_Pete]

Yes it is the lack of phase indication with a one sided null balance meter that is the real issue with these small meters.
Pete.

[Skywave]

I found that it is very helpful for discovering the null point, to connect a X10 'scope probe - and thus to a 'scope - to the cathode of the rectifier at the end of the detector amplifier. By choosing the gain of the 'scope's Y-amp., you get a much improved sensitivity (compared to the NULL meter) as the null is approached, plus the exact null point can be easily determined. The rate of change as that null point is approached gives an indication of the Q of the component under measurement.

As regards my earlier remarks about 'coils that won't null', yesterday I was quite surprised to discover that many of those coils would similarly not 'respond' when I hooked them up to my Advance T2 Q-meter. However, I could determine their inductance value on my trusty HP LCR meter, model 4332A, and the indicated values were very close to the values indicated on my Marconi TF868B.

All this reminds me of the old adage about past-times sailors and navigation: they never went to sea with just one compass - always took three. With only one, can you trust it? With three, the chances of all three being at least substantially inaccurate is greatly reduced - almost to the point of zero. Hence, don't have just the one electronic instrument for measurement. Plus, know how to use them - and be aware of their inherent limitations.

Al.

[G0HZU_JMR]

Yes, they aren't easy instruments to use...We had a 2700 at work about 30 years ago and it was largely unloved even in those days. It was at its best checking the tolerance of large handwound inductors. Back in those days we often used sub contractors to wind inductors for us and the 2700 was a useful verification tool for quality control purposes etc.

But it only measures at 1kHz with the internal oscillator and we rarely need to (accurately) measure large inductors these days. By contrast, we also had the classic Marconi TF1245 Q meter. This wasn't a precise instrument by any means but it saw a lot of use for many years. For RF design work it is better to test at the working RF frequency and the TF1245 permits this. The 1245 was still a bit of a fiddly instrument and it took up a lot of bench space but I enjoyed using it.

Here at home I would use the AC voltmeter + sense resistor method to measure large inductance or capacitance (and ESR) at lowish frequencies. With care it will deliver very good results but this does require some experience. The latest ~£100 handheld LCR/ESR meters are fairly decent too. I've always avoided cheapo handheld LCR meters but I will probably buy one of the latest models at some point.

[Skywave]

Although the posts to this thread may be useful to others, since I no longer own this TF2700 - and, moreover, will never be owning another - I have nothing to add to this thread. Therefore I will leave it up to the moderating team to decide whether this thread should be closed.

Al.