Question about Heathkit IG-72

js22

Active Member
My virst venture into tube audio: an audio generator. Because nothing is as sweet as the sound of a sine wave......:crazy:

I found a nice article and schematic online.

20190307_111356.jpg

So far, testing is going well. I disconnected the two 40 uF supply filter caps and tested them for leakage at 400-ish volts. No substantial leaks, so I reconnected them to rectifier, but not to the rest of the circuit. I now have the unit powered on through a variac and plan to let it sit for a while.

20190307_111108.jpg

The voltmeter is connected to the output of the power supply. So the rectifier is working, the caps haven't exploded, and my hair isn't all pointing straight out. But I have some questions.

1) I brought the variac (gradually) up to about 104 VAC which gave me a DC voltage of 450V. This is the rating of the caps, and also the working voltage of the design. If I plug the unit into the AC outlet directly, I'll be over 450V my a bit. Is this safe? Normally I would *never* run caps at their rated voltage.

2) What to do if I buy new caps? Can I still go with 450V ?

3) Looking ahead, there is a cap in the upper left part of the circuit, 16 uF/150VDC. It's part of the oscillator. It seems to me that if the 6AU6 tube were missing or inoperable, the cap would gradually charge up to the full 450V of the supply; it's simply in series with about 121k ohms. That seems pretty dangerous. Is that just the way it is, or am I missing something?
 
No load on the power supply, so it'll be high. It'll drop a bit under load.

3) The cap won't reach 450V, its leakage will increase above 150V and it won't get much higher than 150. The resistor would be dissipating about 3/4W at that point, not so good.
 
I'd say replace all the electrolytic caps. New caps are cheap, better quality today, much more compact than the originals, and will be more reliable. Even if you can reform the old caps, there's no guarantee that they will not fail right away. There's plenty of room in that chassis to put a tag strip with replacement capacitors. Desolder the old can cap connections (assuming it's a can with multiple sections) and take it out of the circuit entirely.

How old is this unit? Line voltage today is higher than it was back then, and you don't want to exceed the rating of the caps. Tom Bavis is correct that the voltage will drop with the load of the tubes connected, but I would not assume that it's going to be ok. Measure! If you need to drop the B+ a little, you can add a resistor between the 6X4 and the first capacitor.
 
Last edited:
Thanks! I'm new to tubes, so I haven't developed the instincts for understanding these circuits.

After testing the 2 remaining caps, I put it all back together, DeOxit-ed it, crossed my fingers, and turned it on.

No oscillation, except for short bursts. Then I checked the DC power supply and it was way lower than before. I had forgotten to turn up the variac from the "resting" level I used earlier. Measured VAC in and VDC, and sure enough at 110 VAC the DC supply was well below 450, so no worries there.

Oh, and my scope was full of sine waves.:banana:


Here are some quick and dirty results using the digital scope's FFT function.

First, with no input connected:
20190307_152449.jpg

Next with a modern (Instek) function generator:

20190307_152511.jpg

Hard to see (my photographer sucks...) but a pretty noticeable hump at 3f.


Finally, with my $10 1960's IG-72:

20190307_152531.jpg

Just like Colonel Klink, "I see Nothing!"
 
Last edited:
I'd say replace all the electrolytic caps. New caps are cheap, better quality today, much more compact than the originals, and will be more reliable. Even if you can reform the old caps, there's no guarantee that they will not fail right away. There's plenty of room in that chassis to put a tag strip with replacement capacitors. Desolder the old can cap connections (assuming it's a can with multiple sections) and take it out of the circuit entirely.

Yeah, I've already put the new caps in my shopping cart at Digikey, but that's not the same thing as playing with my new toy RIGHT NOW!
 
I'd say replace all the electrolytic caps. New caps are cheap, better quality today, much more compact than the originals, and will be more reliable. Even if you can reform the old caps, there's no guarantee that they will not fail right away. There's plenty of room in that chassis to put a tag strip with replacement capacitors. Desolder the old can cap connections (assuming it's a can with multiple sections) and take it out of the circuit entirely.

Exactly!

Whenever a filter capacitor shots out the rectifier arcs from overcurrent and is destroyed, and that burns out the B+ winding from over-current because, being unfused, the main transformer fuse is set for the heater current for the tubes and is thus far to high to protect the B+ winding.

This is why I suggest adding low-current fuses to the B+ winding, right before the rectifier, and then a fuse right after the rectifier in case the filter capacitor shorts.

How old is this unit? Line voltage today is higher than it was back then, and you don't want to exceed the rating of the caps. Tom Bavis is correct that the voltage will drop with the load of the tubes connected, but I would not assume that it's going to be ok. Measure! If you need to drop the B+ a little, you can add a resistor between the 6X4 and the first capacitor.

Another good point.

It's better to add a bucking transformer than use a dropping resistor as this removes the heat issue and provides better regulation, as it is not based upon current draw.

If one is set on using dissipating voltage in a resistor, far better to add a voltage divider after the filter capacitor. A series dropping resistor may affect the charging cycle of the capacitor, and can thereby increase the load on the rectifier. Also, a series resistor must be sized based upon current draw which may not necessarily be constant so the voltage may fluctuate.
 
Thanks, Retrovert.

I like the fusing idea. Cheap and reassuring.

Also I saw online a guy replaced the 5k 20W resistor with a higher-wattage chassis mount version. That sucker is less than One-Mississippi hot! This instrument doesn't see the hours of use that a TV or stereo would, but that much undissipated heat seems inelegant.

It's funny, this little venture into the dangerous of the tube era makes me realize why people used to be so terrified of electricity - because it could so easily kill you or at least destroy your home. It "shocks" me how complacent young people ate these days. But let's not open that can of worms....
 
Very cool. Heathkit FTW!

I built a couple of modest Heathkits as a kid (I remember the VOM in a stylish brown plastic carry case.) But I devoured their catalogs. Seriously- hours at at time. In my world view, there was a special circle of Heaven where everyone got to use their Heathkit tools to build their very own Heathkit stereo and TV.

On a sad note, I went back to the thrift store today to check on the audio generator's formidable big brother, the IG-42 lab generator. At a spendy price of $20, I thought it would survive until half price kicks in this Saturday. Sadly, that wasn't the case. I hope it found a good home.
 
I have an IG-37 in the queue for restoration - after I learn more about FM and alignment! Restoring the unit should be straightforward.

More pictures of your sine generator would be welcome.
 
I like the fusing idea. Cheap and reassuring.

I don't know why everyone isn't adding this to tube gear to protect the power transformer and rectifier tube. I wouldn't put a fuse in the signal path, because it causes distortion, but for sure, sprinkle fuses about the power supply where it prevents expensive failures.

I suggest using the lowest value fuse that will work. Fuses do not instantly fire, so there is some headroom and that should be reduced as much as possible. For an instrument the demand is going to be small, so you may be limited to a fuse which is something like 1/8 Amp.

Also I saw online a guy replaced the 5k 20W resistor with a higher-wattage chassis mount version. That sucker is less than One-Mississippi hot! This instrument doesn't see the hours of use that a TV or stereo would, but that much undissipated heat seems inelegant.

Shortens the lifespan of electrolytic capacitors, too. The lifespan exponentially drops as temperature increases.

Manufacturing cost drove the design of most equipment, and components were expensive. The ones that were not as focused on cost at the expense of quality, such as HP and Tek, lost sales at the low end to competitors who made such tradeoffs.

We have wonderful components with amazing properties for trivial cost. Certainly nothing like that was available back in the days when this equipment was designed and fabricated.

It's funny, this little venture into the dangerous of the tube era makes me realize why people used to be so terrified of electricity - because it could so easily kill you or at least destroy your home. It "shocks" me how complacent young people ate these days. But let's not open that can of worms....

That rapidly turns into the geezer argument, "Eh, the kids these days with their social media and cell phones. Never used equipment with voltage high enough to kill you, didn't fight the Nazis or the commies like we did, and never tried to crush a steel beer can. Why, back in my day..."
 
I have an IG-37 in the queue for restoration - after I learn more about FM and alignment! Restoring the unit should be straightforward.

Not to dissuade you from a rebuild, but when I was looking for a FM generator for alignment I looked at a bunch of the tube and solid state units. The tube ones require several hours to stabilize and the accuracy was not as good as the solid state units. I don't know if this matters as the tube tuner being aligned has its own set of inaccuracies and drift, particularly in the MPX circuitry.

I remember Conrad had posted some analysis of the accuracy of various FM generators. I think this is it:
 
I finally went through check-out at Digikey and 2 days later got my caps. It was probably overkill but I decided to "stuff the can" for the PS filter caps. My first time at this. Stinky!

Here's how it came out:
20190327_180159.jpg

Not much to see. I used a little bit of perfboard for strain relief on the caps and to keep the leads from shorting out.

Seems like most caps are radial these days rather than axial. So for one cap that was hanging mid-air over a long span, I again whittled a bit of perfboard and hot-glued it diving-board style to the inductor:

20190327_180133.jpg

Also visible to the left are the 4 10W resistors I used to replace the single 20W resistor. I made a bracket to clamp them (firmly, but not too hard) to the aluminum chassis. It's not great thermal contact, but these resistors are not meant for a heat sink. Right now they're dry, but I'll probably add some thermal grease.

So I powered it up and..... Nothing.

The no sparks/smoke/electrocution part of this is good news indeed. The power supply is working and there is an output, but its DC. I think I need to mess with the "oscillator control" pot now that the capacitance had changed. I'll try that tomorrow.

TBH, I had completely forgotten about Retrovert's suggestion about adding fuses. Glad I revisited this thread. I will do that soon.
 
Nice work on the clamp for the resistors. The multiple resistors likely runs cooler than the original.

The funny thing about oscillators is how easily one may stop the oscillation. Conversely, the funny thing about amplifiers is how easily one may cause oscillation. Great mysteries.

In the future when you turn on equipment after a rebuild you might want to use a dim bulb tester to limit the current. That will someday spare you considerable damage because of a short which takes out a rectifier or a transformer winding.
 
Another funny thing about oscillators: if you omit a wire and break the feedback path they get all pouty and just sit there doing nothing. DAMHIKT. (It was pretty obvious in the previous pic of the "diving board" capacitor with only one wire!)

Anyway, it gave me a good chance to find out how scratchy the osc control and meter cal pots were. I gave those a nice clean and lube and now they're like butter.

The frequencies are somewhat (20-30%) off from the rotary switch values. I haven't decided if it's worth it to R&R all the R's and C's to get it "in tune". Probably not.

Here's a pic of the IG-72 in action, pretending to do something useful:

20190328_135644.jpg
 
Another funny thing about oscillators: if you omit a wire and break the feedback path they get all pouty and just sit there doing nothing.

(sputters) But feedback is evil and must be removed from every circuit! Didn't you learn that in Audiophilia 101?

Anyway, it gave me a good chance to find out how scratchy the osc control and meter cal pots were. I gave those a nice clean and lube and now they're like butter.

Always a nice outcome.

Did you polish the contacts in the rotary switches? Heathkits rotary switches, like any other silver-contact switch after this many years, typically have the contacts covered in oxides and sulfides. Metal polish on q-tip or cotton pad is an easy way to remove it. I usually wipe off the residue using a cotton pad with limonene. That should be compatible with the phenolic switch material.

The frequencies are somewhat (20-30%) off from the rotary switch values. I haven't decided if it's worth it to R&R all the R's and C's to get it "in tune". Probably not.

You've gotten this far. How many resistors and capacitors need to be replaced? With 1% metal film and 1% PP it should be spot on, likely better than new.

I just picked up a Cornell-Dubilier BF-50 capacitor tester. It requires replacement of all of the resistors and capacitors which should greatly improve the accuracy.

I expect that I won't need to hand-select for the bridge, as off-the-shelf 1% parts are likely better matched than what was hand-matched in 1939.

Here's a pic of the IG-72 in action, pretending to do something useful:

That's a very clean unit, inside and out.
 
Did you polish the contacts in the rotary switches? Heathkits rotary switches, like any other silver-contact switch after this many years, typically have the contacts covered in oxides and sulfides. Metal polish on q-tip or cotton pad is an easy way to remove it. I usually wipe off the residue using a cotton pad with limonene. That should be compatible with the phenolic switch material.

Those wafer switch contacts always look so flimsy and fragile to me. I've always been afraid to do anything more than DeOxit.

So what you are suggesting is to first use metal polish to clean it up, then the limonene to get rid of the polish? Do you just polish the moving disc contact w/ a Q-tip or is there a way to polish inside the fixed contacts? Do you just exercise the switch with the polish applied?
 
Those wafer switch contacts always look so flimsy and fragile to me. I've always been afraid to do anything more than DeOxit.

So what you are suggesting is to first use metal polish to clean it up, then the limonene to get rid of the polish? Do you just polish the moving disc contact w/ a Q-tip or is there a way to polish inside the fixed contacts? Do you just exercise the switch with the polish applied?

The corrosion layer is hard enough that a conventional contact cleaner like DeOxIt may not remove all of it.

I've done polished the contacts in open formaldehyde-phenolic (brown) wafer switches with great care and without much force. This removes noise and scratchy behavior.

Contacts coated in corrosion are often noisy or scratchy because metal oxides and sulfides are semiconductors. The first rectifiers, in fact, were made from metal oxides. Copper, aluminum, selenium, etc. So having that on switches, potentiometer wipers, jacks and plugs, etc. can affect functioning, particularly as it is frequency dependent. But it also happens down at low frequencies where one would not expect it.

I have a NuTone doorbell using a motor-driven rotary switch as a sequencer to control the solenoid firing. The switch is four poles, so each switch position controls a separate solenoid. The 1950s approach to an embedded controller, if you will. It simply would not properly function because of corrosion and dirt. The contacts were black/brown which is silver oxide and sulfide, and it was dusty. After cleaning and lubricating the switch smoothly transitioned and the solenoids properly fired every time. It should last another 65 years, which is certainly longer than I will. I must disassemble the Telechron movement as the animal-fat grease has congealed, and it was only by working solvent and silicone into the opening at the shaft that I was able to get it to work. The clock still doesn't reliably work. The interwebs describes how to disassemble those soldered cases using a torch, but I haven't yet done that.

The switch contacts are visible when not covered by the wiper and thus easily accessed for polishing. The wiper, as you concluded, is polished by having it move over the contacts which contain polish. In only rare instances, almost not quite nearly never, am I willing to disassemble the switch. Afterwards I carefully wash away the polish and removed corrosion. DeOxIt will do that. So will distilled water. Or limone, but I tend to use a Q-tip to apply that. I do not suggest or recommend saturating switches in solvents.

When finished with the polishing and cleaning don't forget to re-lubricate the switch with a dollop of silicone grease. I've used the SuperLube which has PTFE (tradename Teflon). This greatly reduces the friction and prolongs the lifespan of both the mechanical bearings and surfaces in the switch, but also the contacts.
 
I am in the process of restoring an IG-72 which the original builder had built with diode rectification and had added a 7500ohm 20W resistor in the B+ line going to the filter caps.

All of the electrolytics were leaky and were replaced. Rather than install new filter caps under the chassis, I removed the dual 40uF can capacitor, cut the can off and then re-used the base, including the metal "ring" at the base of the can. I used a pair of 47uF @ 450v caps.It works very well and the caps are above the chassis and out of the way.

First tests show the B+ at ~ 413Vdc with all tubes in place and active, which is far over the voltage rating for the 6CL6 tube. The secondary B+ feeding the 6AU6 is about 240Vdc which while a little higher than the manual shows, is still within the ratings of the tube.

I'm going to let the unit burn-in for a few hours to form all the new electrolytics which include the 22uF @ 350Vdc replacement for C-8 and the 22uF @ 250Vdc replacement for C-7, then I'll try the alignment and calibration routines according to the manual...

I might even take some pictures and post them here when I'm done....Jordan in Calgary....
 
Back
Top Bottom