Function generator output level drifting

[David1]

The output level on my BK Precision 4017A function generator is drifting.

For example, if I set the output level to 10.0V on a 2kHz sine wave, driving a variety of simple half-wave rectifier and diode clamp circuits, the function generator output level will hold fairly steady for perhaps 10 or 15 minutes then drift down to around 6 volts or up to around 15 volts over a period of a couple of minutes, then settle at the new level for a while.

Any suggestions what might be wrong please e.g. dodgy potentiometer, capacitor, or transistor, or are there just too many possibilities to diagnose? I have read the manuals, Googled FG trouble shooting (a lot on frequencies but nothing so far on output level) and emailed BK Precision support but no reply and I guess they are closed now for the weekend.

[Armadillo]

Are you describing a offset level shifting or amplitude level shifting?

I would scope power supply voltage ripples from all the power supply outputs and identify which one is faulty.

Also clean all the contacts. Replace all faulty electrolytic capacitors.

[DC1MC]

  Holly cow, I was complaining about FY6600 and somebody paid in excess of 400USD for this POS ?!?!

This thread summarizes it:
https://www.eevblog.com/forum/testgear/should-i-buy-a-bk-precision-4017a-function-generator/

Anyway, besides you, not many people have seen the schematics of this thing, and such an ENORMOUS drift means something it's seriously screwed (bent on a elco/tantalum breakdown)  but will be a very easy to measure screw-up, did this thing has an external power supply, have you checked it ?

Post the schematics so we can have a look, you won't get any manufacturer support besides "shipping it to us if it's in warranty time".

 Good luck,
 DC1MC

[David1]

Thanks Armadillo and DC1MC. Drift is on amplitude level.

Haha, pretty low opinions of BK on that link DC1MC. Spot on though as I am also working through Learning the Art of Electronics. Schematic from BK is attached but it's a poor scan of a 1995 drawing so difficult to see what's what. Power supply is integral, not external.

No warranty as got it second hand (shipping and customs from USA to UK cost as much as the base price).

[DC1MC]

Thanks Armadillo and DC1MC. Drift is on amplitude level.

Haha, pretty low opinions of BK on that link DC1MC. Spot on though as I am also working through Learning the Art of Electronics. Schematic from BK is attached but it's a poor scan of a 1995 drawing so difficult to see what's what. Power supply is integral, not external.

No warranty as got it second hand (shipping and customs from USA to UK cost as much as the base price).

OK, the schematic is bad but usable in an emergency, stick a scope on the Level adjusting potentiometer R300, both on the cursor and on the input and look at the output level, if the level stays fixed on the input, then focus on the cursor's level, if both are stable then the output amplifier is out of question, if the potentiometer input is stable, but cursor fluctuates, check the potentiometer.

But if the input signal amplitude is unstable coming from the oscillator, than you're in a lot of pain,  there are like a brazllion of switches, potentiometers and stuff to clean and verify, first check the +/- 22V power supplies and also +/-15V, you have an RC oscillator there, and it's not unheard of to have amplitude drifts if the power supplies are noisy or drifty, if the voltages seem OK please also use your scope on DC coupling and check the ripple of the PS as well, sometimes, due to component failure, when going from cold to warm, there are drifts and noise shown up. IF you have some analogue meter, put it on each power supply in turn, cool off the generator and let it warm and reproduce the output level drift while checking the power supplies values.


 Cheers,
 DC1MC

[David1]

Many thanks for taking a look and for the pointers DC1MC. It's been happily stable for the last 40 minutes, albeit giving 10V where I would expect 14 or 15V from the position of the output level dial. Need to get up the courage to poke around with the hood up and the engine running... ...carefully picking my way through a brazilian... ...mmm could give that a go!

[amspire]

Looks like it is based around the Maxim MAX038 waveform generator IC.

The good news is that from looking at the circuit, it should be fairly straightforward to locate the problem.

Do you get the same problem with a resistive load around 50 ohms?

[David1]

Thanks amspire. 15V across 50 ohm = 4.5W which would burn up my 0.25W resistors so I haven't tried that. I've got a 25W 10 ohm resistor, but that sounds like a heavy load. Resistors in the diode circuits I'm using are 1k.

[amspire]

Thanks amspire. 15V across 50 ohm = 4.5W which would burn up my 0.25W resistors so I haven't tried that. I've got a 25W 10 ohm resistor, but that sounds like a heavy load. Resistors in the diode circuits I'm using are 1k.

You will find it is a maximum of 10V into the 50 ohms, but that is 2W.

Just try it with the 1K without the diodes. Is it stable?

[David1]

I'll try that, although is has been stable for the past few hours, but I'll leave it running and see what happens.

[amspire]

The diode circuit can be a difficult load. If it is only varying with that diode circuit, then the explanation is purely in the output amplifier stage.

If the problem is earlier, the output will vary with any load.

If the problem was happening initially but it now not happening, then it could be something as simple as a poor contact of the wiper in the level control.

[David1]

Well it does seem to be more stable without the diode, both input to the circuit (i.e. output from the FG) and output from the circuit.

Max amplitude now is 12.2V (i.e. +6.1V, -6.1V) whereas I think it should be around 20V amplitude (i.e max +10V as you mentioned above), and once it gets down to 6.7V amplitude the bottom 1/3 of the output level control has no further effect. So I guess that should be the first thing to check.

[amspire]

The generator has a 50 ohm output impedance. Basically this means that between the output amplifier and the output, there is a 50 ohm series resistance. The way they have actually done it is to design the amplifier with a 5 ohm output impedance and add a 45 ohm resistor in series to the output.

That means that with no load, you should be able to get 20V peak-to-peak out. With a 50 ohm load, you get 10V p-p out.

With a 1K load, you should get 19V p-p out maximum.

If you are actually measuring the RMS voltage instead of peak-to-peak, then 20V p-p is about 7V RMS.

Now the level control is not meant to go to zero. It will go from just under 10% to 100%. If you want lower, you turn on the 20dB output attenuator. The output will then go from under 1% to 10%.

[David1]

It looks beautiful inside, very clean and tidy, but no obvious way to work on the output level potentiometer without substantial dismantling, so I'm going to wait until I've learnt more before tackling that.

Thank you for all your suggestions and narrowing down the problem.

[David Hess]

I have a much older version of this model with a different design.  I am not impressed by this one; the design shown in the schematic is not very good and shows a simple brute force approach.

Measure the output of the DC offset amplifier which is pin 1 of U300A to see if it is drifting.

Inspect the output resistors to see if they have overheated in the past due to an excessive load.

My guess is that you will find that the output transistors are running hot and are damaged.  U301 may be damaged as well.  None of these function generator designs including mine had a fault protected output stage.  I had to change the output resistors and transistors on mine once.

[David1]

So perhaps it was a loose potentiometer lead or similar after all. Took the cover off and gently poked things and it has been dead stable for the past three hours driving diode circuits and with full range to 19.8V. We will see.

Thank you David Hess for your suggestions too.

[David Hess]

So perhaps it was a loose potentiometer lead or similar after all. Took the cover off and gently poked things and it has been dead stable for the past three hours driving diode circuits and with full range to 19.8V. We will see.

Yours is pretty new so I was not expecting a mechanical problem.

[David1]

As a follow up...

BK Precision recommended first checking the output level potentiometer (spot on, it has been stable ever since) and have dug around in their files and provided a component list with zoomed in sections of the schematic. It is a pretty ancient machine (schematic is from 1994), so all in all I say brilliant customer service from BK Precision!

[Armadillo]

Good to hear that it is resolved.

Edit: Loose leads. OK.   

[David1]

Postscript:

Frequency has been unstable for a while, usually fixable with a gentle finger-flick on the coarse control. But a few days ago it just went permanently haywire so I decided to take it apart.

Turns out there's a metal shield that sits immediately above the frequency range buttons separated from their contacts by only a sliver of dried out masking tape! Well, maybe it's transformer insulating film, but pretty flimsy and held in place only by pressure between metal shield and switch contacts. So I've replaced it with a strip of plastic film stuck to the shield with self-amalgamating tape.

Put it back together and it works a treat! I'm so pleased!