HP 608C Signal Generator – Repair Restore

Found this fella in the yard of a prestigious organization. They had many other valuables to take care of and this was low priority/value compared to what they had. Since I am not normal and have this problem of collecting as many as scopes and meters and generators like many others, I took it home.

Now restoring one of these things are not very difficult if you know the basics. Very similar steps to the Tektronix Scopes, but much easy as its a simple instrument compared to a 500 Series oscilloscope with 50-70 Tubes.




Even if you are familiar with electronics, be very careful when you work on something like this, a maximum potential difference of 400V DC is used in this equipment . -165V and +225 is the power rail inside. DO TAKE CARE.

Use test  instruments with right attenuation probes, example, 10:1 probes or HV probes,else you may fry your test instrument & the signal generator.

Now, as always, cleaning, Re-Cap, Test Tubes, Power up, and troubleshoot/fix..ahh..wait !recap ..The thing I hate mostly from the originality of the instrument. I hate to hook up modern HV electrolytic all over. Paper was still okay as there was mostly enough space to tuck the modern mica/film caps in.

I was lucky enough to have a direct conversation with Dennis Monticelli on this capacitor aging issue and got a 10 minute dump from him on the subject, and he explained the whole process of manufacturing and the process to reform. I definitely wanted to try it out and here is the result.

This restoration DOES NOT INVOLVE re-capping electrolytic caps, being HP and they used the best quality components and I was able to reform all electrolytic. {{no such choice on paper/wax/oil}}.

Electrolytic Capacitor Reforming

NOTE: Not all capacitors can not be reformed.

Here is the schematic and method I used to reform the electrolytic caps.

Components in the Cap reforming setup.

Variable High voltage DC Power supply – I used the Heathkit 2717 I restored earlier – 0 – 400V DC.

A Series resistor to limit current – 47K/2W – so max current is <2mA even at 450V. Ideally you need 4-5WW resistor to handle a worse case scenario of a shorted cap, since I am monitoring the whole process, I took a chance.

Two ammeter or micro ammeters  –

A Digital- to read the exact leak current -I used HP 34401A 6.5 digit

An Analog -for better visual Indication of the current

{Good if you can have the analog meter with a switch to change probe polarity, as during discharge the current direction will change. Having a switch helps to avoid changing the probe orientation (+ and -ve) to the meter during discharge cycle}

A DMM to measure voltage across the capacitor being reformed.

A Switch to discharge the capacitor after reforming.

Here is the schematic, its very simple and straight forward with specific instruments I have used to help mapping them with pictures. You can use any equivalent instrument.


Here is the steps I followed.

1. Disconnect the capacitor from the circuit.

2. Wire the leads from the reforming circuit to the capacitor – – ensure correct polarity.

Recommend  that you remove both +ve and -ve leads from the capacitor from the circuit it lives in.

{{WARNING : if you are reforming caps on -ve power supply rail, the +ve of the capacitor “could” be grounded to the chassis {{in most cases}}. So the equipment chassis  will hold B+ from the REFORMING power supply. BE CAREFUL. If unsure, remove the cap off from the equipment and reform}}. Same applicable for CANs with unprotected body cap. Usually the body of the can capacitor is wired to -ve of the capacitor, so take care.

3. Set ammeter for 1mA range – 0 – 1mA – or 0-1000µA.

Update from  Denny -9/13/18 – “The lower the reforming current the slower the oxide regrowth and the higher the quality of the oxide. 1mA max is just fine for routine reforming jobs, though”. So if you really have time and wants the caps in  best condition, may be you want to lower the current (100K or 220K/2W resistor instead of 47K).

4. Slowly increase the voltage output from the HV variable power supply, until the current reaches 1mA, and wait for it to come down to less than <300uA-500uA

{The actual allowed leakage current may vary according to the capacitor value and voltage its specified for, check general datasheets for typical numbers }

5. Watch and confirm that the voltage across the Cap is more or less equal to Power supply output – actual voltage across cap will be –

Power supply out – (leak current x 47k)

6. Increase the Power supply voltage a again, limiting max current to ~1mA, wait for the cap to charge and reform and the current to drop to less than <300uA-500uA

7. You may find the current in the cap fluctuating during the reforming cycle.

8. Repeat steps 4-6 in steps. until the voltage applied across the capacitor is close to its rated voltage –  I did 430V on a 450V rated cap.

{{even though Heathkit 2717-A is a 0-400V regulated variable power supply, you can tweak it to go to ~ 430V. Check the documentation and adjust 400V out adjust Preset to a bit on the higher side}}

9. Once I had like 430Vs on a 450V Capacitor, I left it for a while, until the leakage current was reduced drastically, I could get 180uA leakage on a 20uF/450V Capacitor.

10. Change the polarity on the analog multi-meter, set it to a higher current range, and move the “Charge-Discharge” switch to Discharge.

10. Reduce the power supply output to zero.

11. Wait until the voltage across the cap becomes zero. Short the terminals and Leave it for few minutes and keep the short across the terminals while you try to reinstall it. Else a voltage can develop across it. This is a known effect, called dielectric absorbtion \ or recovery voltage phenomenon, google for details.

It may take anywhere from 15 minutes to 1 hour or more to reform, depending on how bad the cap is and some of them could be beyond rescue. the leakage current will tell you the progress/condition.

To compare, if you  look at leakage current specification for a modern(2018)  Nichicon  electrolytic capacitor series -6.8 to 680µF/160 to 500V.

So if we take a 22uF/450V Nichicon, the leakage current is specified as

“After 1 minute’s application of rated voltage at 20°C, leakage current is not more than 0.04CV+100 (µA) ”

which is =(0 .04x 22µF x 450) + 100µA = 396µA+100µA  = 496µA ~ 500µA.

The leakage current I recorded (on HP 34401A) with the reformed caps were ~ 180µA.-220µA, well within the margin or modern caps.

I wish I could show you how to wire things up and setup the cap reforming, however, unfortunately I can only show you the other side, what NOT to do.

how NOT TO  setup for  reforming Caps.


At the time of this writing, I assume,  I was not electrocuted in this mess of cap reforming circuit, I am still alive and its not my ghost who is blogging this.

In case you heard that I got fried in this mess, that I reformed the caps and deformed me, just FYI,  hell does have Internet access, and sorry none of bad bold girls who died young are here to entertain you.

Now Back to restoration. The entire process of cleaning and restoring is documented here in a video –

In case you like pictures, here is a quick summary of the process

1. Cleaning

2. Test/Replace bumble bees.


3. Electrolytic Reforming

Three tubes were bad, 6080 from power supply, one 6CL6 from RF limiter and one 6BQ7 from the modulation limiter.

Modulation indicator was not working due to weak 6BQ7 , 6080 in the power supply was leaky (gas) , and 6CL6 was way too weak and output level was low.

And the Final output,  @ 413Mhz.

Screen Shot 2018-09-07 at 4.36.05 PM



For more Crazy Stuff like this, Back to Home Page
















7 thoughts on “HP 608C Signal Generator – Repair Restore

  1. Hey Raj! Denny at CHRS posted this article to some of us. Nice restoration! I check leakage and reform as well, in audio amplifiers. My rule of thumb is .5mA per 10uF @ 450V leakage. Over that and the cap will probably head south soon. Let’s chat more when we see each other at Radio Central. Cheers! -Kent


  2. Great repair! I recently obtained a 608C with all working tubes, but still fully equipped with bumblebee capacitors. What caps did you use to replace the bumblebees? Unfortunately, I can’t reform the power supply caps and will replace them with the appropriate modern electrolytic caps.


      1. Another question: would it be permissible for me to use the generator without reforming or replacing the old electrolytic caps? It’s quite expensive and time-consuming to do so – if possible I would prefer to forgo that. Thanks again!


      2. Depends, if the caps are bad, they can blow the rectifier. May be power it up on a variac, say like at 30V and keep it for a while(10 mints or so) , as that also aids reforming, check and confirm the voltage across main rails are to the ratio of mains input, and slowly bring it up to say 60V or so ..leave it for a while..and so on until you can confirm that the caps are holding on, and then move to full 120V. This is the safe bet.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s