Quadrature Oscillator Power Supply

On my old Thorens HiFi deck the motor is driven by a dual op-amp w direct coupled bipolar amp output quadrature oscillator (Wien bridge followed by derivation stage). Input is 2x 12v AC and output on each wave is 6v into 27 Ohms.

It's broke.

I have the schematic, and can send it, but the parts are old eg op-amps are LM-3709C and trannies are 2N2218A/2N2914.

I think it's the semiconductors, because the unit kind of worked until I tweaked it!

Would someone be able to suggest drop in replacements for the Op-Amps and trannies (can replace resistors/caps if requ'd). These must be obtainable from Maplin or other on-line supplier preferably in the UK. I can get datasheets but don't have the knowledge to recalculate bias resistors etc. The other point is that the lamp is a spec which is no longer obtainable so the Wien gain adjusting pots may need to be replaced.

Thanks in antiicipation,



Quadrico
Bucks, UK

Please post the schematic.

Quad Oscillator PSU

Thanks for your reply. The schematic is attached - both an original scan (pretty illegible) and a redrawn version.

In the redrawn version the 2x 500Ohm presets are in fact grouped in threes and switched with the Wien Bridge pots - ie for each frequency there is a separate gain adjustment on both op-amps.

The base frequency for 33 rpm is +/- 40 Hz but in fact probably more like 36Hz.

I see that this week Maplin are offering the LM741Cn as a direct drop in replacement for the LM3709C in most applications. I think I will just write off and get 4! (and some sockets).

Do you agree?

And for the trannies I am thinking BD 140/BD 139?

Would you add a voltage regulator(s) to the PSU?

Do you think the original caps are likely to be OK - all look OK though the PSU rails are not giving equal volts - obviously I actually need to start decoupling bits and testing from the PSU onwards. The issue is they are all +/- 5% which I have not found on the net.

I have a scope running off my PC soundcard for waveform and a decent mutlimeter (Fluke) - so ready to go!

Thanks again,

R



Tks

The datasheet that Maplin posts for LM741 does not have a 14 pin package option. I don't think the 741 has ever been offered in that package.

It has but that was a long time ago.

Well I had thought that since pins 1 2 13 14 on the 709 were nc it would be a drop in to a socket - but it won't. The 741 has only one frequency compensation input and the 709 has two both of which are used.

Back to the drawing board?

Tks,

R

Though looking more closely at the datasheet schematic it may be that the 741 has internal frequency compensation in the form of C1 and R7/R8. This is getting rather involved!

Any ideas?

R

The 741 is internally compensated, and has no compensation pins, as I think you have discovered. Since the oscillators run at low frequency (I think), the bandwidth of a 741 would be adequate, although there are much better choices available (e.g., TL071).
The big issue is packaging. Did you read my previous post? The 709 was in a 14 pin package. I don't know where you will find a drop-in replacement in a 14 pin package. Maybe someome else does. You could possibly make a little adapter to carry a different package.
EDIT: I see on the National 741 datasheet that it is available (?) in a ceramic package, which is probably pretty expensive, if you can even get it.

Perhaps you ought to explain what you did to 'destroy it'?, you most probably don't need to go round blindly changing semiconductors (it might not help, and may make it worse) - fault find, find what's faulty, and repair or replace that.

I just happen to have a diagram of the old 14 pin version of the 741, and its pin connections are the same a the 709 with the exception that Roff mentioned of internal frequency compensation so those pins have NC.
This may be a solution. The 8 pin version of the 741 has the following pin assignments:
1 off set null
2 - input
3 + input
4 V-
5 off set null
6 out
7 V+
8 NC
Here is a possible solution to the existing problem.
If you plug a 8 pin LM 741 replacement into a existing 14 pin socket with the 8 pin IC pin 1 into the pin 3 of the 14 pin socket everything would be ok.
You could remove the frequency compensation parts if you wanted to.

Wow, nice observation, k7elp60! I still think I might switch to TL071, but it probably doesn't matter.
I think Nigel's recommendation should be followed first.

k7elp60's idea is the same as mine - not so well explained. Thanks.

Nigel is of course correct but here's the story: originally the board was 'sort of' working though it was very difficult to balance the two outputs. The motor actually seemed to run best with the sine wave at ~7v at which the cos wave was only about 2.8v

However 'best' in this context still meant only at 45 rpm and with some cogging. The motor is good as I have powered it from a .wav file through a power amp and at 6v it runs fine.

This of course suggests firstly that some of the switchable components were the problem. However, the PSU rails were also not equal at +12v and -9v.

During the course of some tweaking by moving the pots something gave out and now I can't get a decent output at all.

Since starting this I have acquired the freeware Scope program and can get some waveforms. My problem is I think that if one op-amp is blown it will load the other and if the trannies are blown they will load the output of the op-amp, in ways that I would not be able to interpret.

So I though it best to do a clean sweep. If people still think it best to try the incremental route I would be glad of advice where to start. It seems obvious (but that doesn't mean it's right) to separate PSU and op-amps, but not sure where to go from there.

Thanks for the ideas, guys.

R

I would suggest that as it died when you tweakd a pot, the pot may well be duff? - old pots not only tend to go faulty, they also get very fragile, and it may have broken when you adjusted it.

But the circuit is simple enough, a free running wein oscillator (RHS) and an invert (LHS) - both have a simple pair of transistors to increase their current sourcing capacity.

The 500 ohm pots set the gain of each stage, set the oscillator one first, so the output is high but below clipping (unless you can find proper setting instructions?). Set the inverting amplifier one to give the same size output.

If the amplifier feedback pot is maladjusted, the oscilaltor won't run at all, so it could that that's faulty.

Some more investigation yields the following:

1

Some more investigation yields the following

1 Thankfully the pots are all OK

2 The PSU is OK giving + and - 18v no load and no ripple

3 Isolating the Wien bridge circuit gives the following results - first trace is from the putput of the current ampliying transistors and the p-p volts here are 2.9v as measured by Fluke digital MM. Second trace is from the output of the OpAmp and is approx 9v p-p.

What next?

Thanks

R

In the above post the traces are transposed - Trace 1 is the spiky one and Trace 2 the square one. Tks

R

I should also have mentioned that I can get a sine wave from the Wien bridge but only at 0.2v p-p at the transistor output and unmeasurably low at the op-amp. The other point to consider is that the gain is set for the combination as a whole - ie the feedback to the opamp is from the transistor output, not the opamp output.

That's all on this subject for tonight!

Tks,
R

9v p-p? I can't reconcile that with the scale factor on your photo.

That's right. Because the scope is basically using my PC's sound card, I have bridged the line input with a 10k pot. Thus the amplitude of the wave on the scope screen is dependent both on the pot setting and the sensitivity of the scope setting.

Brgds,

R