BK Precision 4010 2MHz Function Generator repair; Need help!

[SittingBear]

SOLVED
See below for solution.

Hello EEVBlog,
I just recently purchased a "never been used" BK Precision 4010 Function Generator! It isn't the 4010a, but from what I can tell it seems identical aside from the exterior.

Link for the lazy: http://www.bkprecision.com/products/signal-generators/4010A-2-mhz-function-generator.html
My issues are:

• Datasheet states 20Vpp (10Vpp at 50? Output), but I'm getting somewhere around 400mVpp at max dial output (if I remember correctly).
• All frequency ranges, and when none is selected, the output has a constant ~120Hz 'carrier' wave which adds and distorts selected frequencies.
• Quite a lot of noise on the higher frequency range (>200kHz).
• Square wave is not very square at all, there's some rogue capacitance somewhere.
• PWM adjustment dramatically lowers the output's offset even when the desired offset is none. To be more specific, higher duty cycle offsets the entire waveform into the negatives.
• Can't find a circuit schematic, but 100% through hole components make it easier to just visually trace.

I'm not quite sure where to begin on this circuit as to what could be the issue.

I will post pictures of the output on the scope tomorrow, but for now I will attach pictures of the circuit board.

Please note: I disconnected and measured two resistors which looked burnt, but ended up being good as far as measured vs. color code goes. I have also tried fiddling with the adjustment potentiometers which did fix my issue of the frequency dial not being calibrated. However, all that did not affect the output were reset to their original states.




From left to right (with datasheet links): CA324E, MAX038CPP, MC75108P, SN74LS04, CD4081BE, and LM1458N




From left to right (TO-93 transistors): 2N2905, and 2N2219A

[tautech]

"Smells" like 2 issues.
PSU needs to be checked for excessive ripple.
If it really has never been used, the ripple may reduce as the Ecaps reform with the power on.

If the output is not SC protected, the output amp could be damaged.
Trace the output waveform before the output stages and check for "purity" and loss of gain near the output.

[SittingBear]

"Smells" like 2 issues.
PSU needs to be checked for excessive ripple.

The PSU seems like a very basic transformer, full-wave rectifier (possibly where the 120Hz comes in?), with some high output caps. I'll check it regardless.

If it really has never been used, the ripple may reduce as the Ecaps reform with the power on.

Sorry, I'm not sure what this means. I think I somewhat remember / understand: my guess is that when capacitors haven't been used in a long time people recommend applying a low voltage to them so they may 'rebuild' the dielectric.

If the output is not SC protected, the output amp could be damaged.

I was planning on replacing the op amp, because I have some very identical ones with exact same pin-out and similar ranges.

Trace the output waveform before the output stages and check for "purity" and loss of gain near the output.

I'll trace the waveform along it's path when I get time tomorrow, thank you so much for the help! 

[tautech]

Quote from: tautech on Today at 02:46:40 PM
If it really has never been used, the ripple may reduce as the Ecaps reform with the power on.
Sorry, I'm not sure what this means. I think I somewhat remember / understand: my guess is that when capacitors haven't been used in a long time people recommend applying a low voltage to them so they may 'rebuild' the dielectric.

Exactly

Seems from what you have posted you have your "head around it" 

But you really need a manual to properly trace and fault find.
I will have a hunt......

[David Hess]

If the output is not SC protected, the output amp could be damaged.

I was planning on replacing the op amp, because I have some very identical ones with exact same pin-out and similar ranges.

The usual failure is the discrete output stage transistors.

[SittingBear]

Quote from: tautech on Today at 02:46:40 PM
If it really has never been used, the ripple may reduce as the Ecaps reform with the power on.
Sorry, I'm not sure what this means. I think I somewhat remember / understand: my guess is that when capacitors haven't been used in a long time people recommend applying a low voltage to them so they may 'rebuild' the dielectric.

Exactly

Seems from what you have posted you have your "head around it" 

But you really need a manual to properly trace and fault find.
I will have a hunt......

It seems that the output of the MAX038CPP is fine, but being lost somewhere along the circuit. It ties into the base of a 2N3906 found in the circuit picture

(Far right transistor, TO-92 package, in the middle).

The usual failure is the discrete output stage transistors.

Which type might those be and where might I find them along the output? Thanks

[David Hess]

The usual failure is the discrete output stage transistors.

Which type might those be and where might I find them along the output? Thanks

I think your last picture shows them.  They are the ones in the TO-39 metal cans with the black finned heat sinks attached.  The output stage is a typical low power class AB emitter follower design.  There should be a couple of 2 or 3 watt power resistors with a value of 4.7 ohms or so joining the emitters and those resistors are usually burnt if the output transistors shorted.  Everybody seems to have copied the same basic 20 year old function generator design up until recently.

You should be able to read the part number off of the parts themselves but 2N3503 and 2N3725 are typical.  They are roughly equivalent to a 2N4401 and 2N4403 but the metal TO-39 package allows heat sinks to be attached for better power dissipation.  I am not sure a TO-92 packaged transistor with a heat sink would be good enough so I would use something like a TO-126 packaged KSA1220 and KSC2690 as replacements.

If I was designing one of these from scratch, I would use multiple 2N4401 and 2N4403 class transistors or better in parallel for the output stage if I did not go with surface mount.

[SittingBear]

I think your last picture shows them.  They are the ones in the TO-39 metal cans with the black finned heat sinks attached.  The output stage is a typical low power class AB emitter follower design.  There should be a couple of 2 or 3 watt power resistors with a value of 4.7 ohms or so joining the emitters and those resistors are usually burnt if the output transistors shorted.  Everybody seems to have copied the same basic 20 year old function generator design up until recently.

You should be able to read the part number off of the parts themselves but 2N3503 and 2N3725 are typical.  They are roughly equivalent to a 2N4401 and 2N4403 but the metal TO-39 package allows heat sinks to be attached for better power dissipation.  I am not sure a TO-92 packaged transistor with a heat sink would be good enough so I would use something like a TO-126 packaged KSA1220 and KSC2690 as replacements.

If I was designing one of these from scratch, I would use multiple 2N4401 and 2N4403 class transistors or better in parallel for the output stage if I did not go with surface mount.

The transistor 'can' with no marking on the top surface is a 2N2219A and the other is a 2N2905. The resistors surrounding them seem to be good. I may remove them from the circuit board and test them with a multimeter for continuity and possibly the hFe / beta of the transistors.

Regardless, I ordered some replacements for each transistor for a relatively cheap price so I may try that.


EDIT: Below are a couple of more naked close-ups of the circuit board in question.

[SeanB]

Seeing as the heatsink on the 2N2905 had been knocked down it likely shorted out the transistor, and probably killed the 2N2219 on the other side. You probably will have to replace both that and the 3 smaller driver transistors, and probably R213 has been damaged with the track going open circuit. As well check that the sleeve on C204 is not shorted out to the 2N2219 transistor.

The smaller transistors are likely 2N2907 and 2N2222, so you can change them with the 2N2905 and the 2N2219 respectively if you have multiple spares. Preferably get the A versions as they are tighter spec devices, generally closer to the middle gain wise and lower leakage.

[SittingBear]

Seeing as the heatsink on the 2N2905 had been knocked down it likely shorted out the transistor, and probably killed the 2N2219 on the other side. You probably will have to replace both that and the 3 smaller driver transistors, and probably R213 has been damaged with the track going open circuit. As well check that the sleeve on C204 is not shorted out to the 2N2219 transistor.

The smaller transistors are likely 2N2907 and 2N2222, so you can change them with the 2N2905 and the 2N2219 respectively if you have multiple spares. Preferably get the A versions as they are tighter spec devices, generally closer to the middle gain wise and lower leakage.

Thanks! I'll make sure to check that out. The commonly used TO-92 transistors are all 2N3904, with about two 2N3906 transistors. I have plenty of those lying around, luckily!

[David Hess]

When I rebuilt my BK Precision 3022 function generator, I replaced the emitter resistors which were discolored from heat with higher wattage metal oxide resistors and added transistor sockets for the transistors that I replaced including the two output transistors.

[SittingBear]

I replaced both can transistors and every other NPN transistor with known goods. The problem is still occurring.

Using my cheap-o IR thermometer, I found out that the LM340T5 regulator is getting quite hot, alongside with the slightly charred transistor even after 30 seconds of the device being powered on. The resistor was slowly climbing beyond 60.0°C while the regulator was about 5.0°C cooler, but also heating up.

I suspect there is some sort of short from power to ground, though it is hard to find in this circuit due to all the low resistance resistors. Some resistors are even close to around 5-ohm. Though I still am not sure what is going on.

The output of the waveform generator chip is still clean and good. I sent in a request to BK Precision for the schematic and calibration manual, with no reply yet.

[SittingBear]

SOLVED

It turns out the fuse blew due to the old transistors and I must have had forgotten to check it. Replaced it now and it works great!