Heathkit IG-1271 Function Generator Overhaul

[AMR Labs]

I like Heathkit stuff, reminds me of the times when you could just walk into your local electronics retailer and walk out the door with a brown box with the heathkit logo and a lot of parts inside to put together and end up with something usefull that you build yourself plus the fun doing it. Another sleepless night... So I like once in a while get something Heathkit off ebay to restore and put on the shelf for eventual use, and the IG-1271 had been on my to get list for a while.

Actually got the SG-1271 version, so this was factory wired, not a kit. No problems with this unit, so just wanted to do the cal procedure to make sure everything is on spec. First step is to do the Lo/Hi dial calibration on the 1K frequency range, which basically entails carefully adjusting both the Low Freq Dial adjust and Symmetry adjustment trimmers to get an even 100Hz triangle wave output at  0.1 dial position, plus set the offset trimmer for even +/- waveform defection, CHECK. Then rotate dial to 10 position and calibrate High Freq Dial adjust trimmer for a 10KHz output, CHECK. And finally set dial back to 1.0 position (where before any further adjustment I'm now reading 833Hz output at this point) and readjust Low Freq Dial adjust trimmer for 1KHz, CHECK. Everything dandy so far, but now when I go to check the frequency output at the 0.1 dial position I do get 225Hz, instead of the expected 100Hz. At this point I am still getting 1KHz at the 1.0 position, and 10KHz at the 10 position, but apparently there is no way to get 100Hz at the 0.1 dial position without disturbing the calibration of the 1.0 position. I've done the calibration steps several times over as indicated in the manual, but always arrive at the same result: 225Hz at the 0.1 dial position, while 1.0 and 10 are still correct. And if I set the 0.1 dial position to be back to 100Hz, the 1.0 position will read 833Hz instead of the expected 1KHz. The 10 position on the dial will still be accurate at 10KHz output as this is set with a separate trimmer adjustment.

Now given that Heathkit has not always exactly been high on accuracy or design perfection (at least not anywhere near a la HP, Tek, etc), I wonder if this error on the 0.1 dial position was considered a "normal", or if there might be an actual still undiscovered problem in this unit, or maybe an error in the write up of the calibration procedure? I've also checked voltages around the main variable oscillator circuit and everything agrees to what is shown in the diagram. Even checked resistance and tracking of the R103 dial pot which is 1K LIN, and to the best of my capabilities it seems to be tracking fine. Function-wise everything else seems to be working fine in the generator, besides the 0.1 dial position frequency error.

So what gives? Before I start what might turn out to be a wild goose chase, anyone out there with an IG-1271, could you please check your output frequency at the 0.1, 1.0 and 10 dial positions (at the 1K frequency range) and report back your findings? Or if anyone have any ideas or suggestions, or maybe also ran into the same calibration problem, I'd appreciate any input.

BTW, since I saw quite a bit of spikes on the +/-15V rails, decided to replace all the 10uF 20V Tantalum caps (seven in total) with fresh 10uF 35V units, but it did not help much to quiet down those rails. Switching spikes exactly correspond to highest and lowest part of sine and triangle waveforms, and rising and falling edges if compared with the square wave output. So I guess these spikes are also part of the original design as well. The main +/- 28V rails looked clean, and the main filter caps still look good, no bulging or any visible defects, so left them alone.

I also noticed both power supply pass regulators (one voltage regulator and a transistor) get extremely hot and are set horizontally against the board with absolutely no heat sinking. I modified both pass devices to be vertical and attached an individual heatsink to each one. They still get hot but at least now the heat is being taken out of the devices to some extent. Same goes for the two main output transistors, which are TO-202 tab devices, they get very hot due to no heat dissipation aid, so I modified two TO-220 heat sinks (new hole) and attached them individually to avoid having to use insulators, as the transistor tabs are each respectively at plus and minus 28V coming straight from the main rails.

Thanks in advance for any input.

[t1d]

And if I set the 0.1 dial position to be back to 100Hz, the 1.0 position will read 833Hz instead of the expected 1KHz. The 10 position on the dial will still be accurate at 10KHz output as this is set with a separate trimmer adjustment.

I don't have any ideas, for you, on this, but I wanted to make comments, about other matters...

BTW, since I saw quite a bit of spikes on the +/-15V rails, decided to replace all the 10uF 20V Tantalum caps (seven in total) with fresh 10uF 35V units, but it did not help much to quiet down those rails.

Good move, on that.

Switching spikes exactly correspond to highest and lowest part of sine and triangle waveforms, and rising and falling edges if compared with the square wave output. So I guess these spikes are also part of the original design as well.

This may be typical, if the IC is an 8038, or XR2206. Even worse, if an LM324.

The main +/- 28V rails looked clean, and the main filter caps still look good, no bulging or any visible defects, so left them alone.

A visual inspection, is a good way to go, for electrolytics. But, they can simply dry out, too. You need to check there ESR.

I also noticed both power supply pass regulators (one voltage regulator and a transistor) get extremely hot and are set horizontally against the board with absolutely no heat sinking. I modified both pass devices to be vertical and attached an individual heatsink to each one. They still get hot but at least now the heat is being taken out of the devices to some extent. Same goes for the two main output transistors, which are TO-202 tab devices, they get very hot due to no heat dissipation aid, so I modified two TO-220 heat sinks (new hole) and attached them individually to avoid having to use insulators, as the transistor tabs are each respectively at plus and minus 28V coming straight from the main rails.

Another smart move, on your part.

[AMR Labs]

t1d, thanks for your input.

Switching spikes exactly correspond to highest and lowest part of sine and triangle waveforms, and rising and falling edges if compared with the square wave output. So I guess these spikes are also part of the original design as well.

This may be typical, if the IC is an 8038, or XR2206. Even worse, if an LM324.

The IG-1271 is not based on or uses any of those chips. The main oscillator (8 discrete transistors) generates a triangular waveform, which is then separately converted to sine by feeding it through a resistor network. The triangle wave is also converted to square wave with a high-speed line receiver SN75107A IC used as a comparator, which I think is where most of the spiking originates. I'm thinking since the IC is soldered in (this unit was factory assembled) and no socket, I will try to disable the square wave section to confirm if the spikes are reduced or go away entirely, and work from there to try and improve the +/-15V filtering.

For reference, a PDF of IG-1271 schematic is here: (not my site)
https://www.rsp-italy.it/Electronics/Kits/_contents/Heathkit/Kits/Heathkit%20IG-1271%20function%20generator.pdf (274KB)

The main +/- 28V rails looked clean, and the main filter caps still look good, no bulging or any visible defects, so left them alone.

A visual inspection, is a good way to go, for electrolytics. But, they can simply dry out, too. You need to check there ESR.

Yes agree, probably should, just to make sure for long term reliability.

[AMR Labs]

Just wanted to bump this thread in case anyone that has access to a IG-1271 or recently acquired one might see it since the last time I got a response some months ago, and possibly assist with my request to just compare a certain aspect of its operation to my unit, which simply involves taking a frequency reading of the output at the 0.1, 1.0 and 10 positions on the dial on any frequency range. My measurements and tests where all done on the 1KHz frequency range, and I verified the problem does not seem to be influenced on which frequency range is ultimately selected.

My unit, which as I mentioned before is the factory build SG-1271 version, seems to be behaving oddly as the signal output at the 1.0 and 10 positions is within spec but at the 0.1 position it is somehow completely wrong and no amount of calibration as indicated by the manual will correct the error. For example on the 1KHz range the output is 1KHz at the 1.0 dial position, and 10KHz at the 10 position, but at the 0.1 dial position I do get 225Hz, instead of the expected 100Hz. Again, this error happens independently of the selected frequency range.

Could anyone please verify the 0.1, 1.0, and 10 dial positions to have output frequencies that are all on spec, or do you see a frequency error as well on the 0.1 dial position? Again, I used the 1KHz range but feel free to use any other range to make your readings.

There are two calibration trimmers, one sets the low end of the dial (at 1.0), the other sets the high end (at 10). One is instructed to set the 1.0 and 10 spots with the respective trimmers, and that works out fine. But the 0.1 spot is then always completely wrong as I just described earlier. I even tried to cal the low end of the dial by setting it to the 0.1 position and then adjust the low end trimmer to get 100Hz, but then the 1.0 position will have an output of 833Hz instead of the expected 1KHz. The 10 position on the dial will still be accurate at 10KHz output as the output frequency at that dial position is ultimately calibrated with the separate high end dial trimmer adjustment.

Before putting more time into trying to correct this really just want to make sure this is not possibly "normal" behavior as Heathkit test equipment is sometimes known to have design shortcomings and/or odd behavior that are unavoidable.

Another possibility that comes to mind is that perhaps my unit came with an odd tracking frequency dial 1K potentiometer, but since it looks to be the original part, and keeping in mind this unit was factory assembled and not sold as a kit, one would think that the factory would have caught the problem while calibrating the unit. Or maybe not.

Thanks for any assistance with requested readings and/or ideas.

[barry14]

I think you are expecting too much from the IG-1271.  According to the specifications in the catalog, the frequency accuracy is 3% of full scale.  Thus at a full scale of 10 kHz (10 on the dial), the expected accuracy is 300 Hz so that your generator is within spec. Note that most of the dial is occupied by the 1 to 10 scale.  The 0.1 to 1 portion is very small and should not be expected to be very accurate.

[AMR Labs]

Maybe, but.... the frequency on both the 1.0 and 10 position is within a few Hz, almost dead on and certainly much better by a long shot than the 3% spec you mention, but still the signal frequency output at the 0.1 dial position is way off spec by about +125% of the expected reading so I will have to disagree that this problem is somehow just related to overall loose instrument specs as you trying to imply. Furthermore, the rotational [angular] distance between the 0.1 and 1.0 positions is the same as the distance between the 1.0 and 2.0 positions, so I see no reason why the lowest scale range should not remain in the same spec range since both end points are in fact in spec. In any case I certainly thank you for taking the time to send in your opinion and input.

I am still hopeful to get someone out  there to make the same measurements I am making to be able to compare the readings on all three dial positions, as that would definitively help to clarify things.

[Brierpath]

Just acquired one of these at a local pawn shop. It has a couple burned resistors(R61 and R47). I will see whether it is alive soon.

[AMR Labs]

I would start by verifying that both +/-26V and regulated +/-15V are present and on spec. The unregulated 26V are usually a bit high, around 30-31V, depending on the AC mains input voltage. R47 usually burns up if power was applied into the generator, most probably with the attenuator set at 0dB. Hopefully Q22 and Q24 are still OK. R61 is just in series with the pilot lamp, and should be an easy fix, but its kind of strange that it burned up.