HP339A distortion analyser

I have managed to make a total donkey of myself and damage a HP339A distortion analyser that I have no documentation for. I inadvertently plugged it into 220V instead of 110V. Naturally the external fuse didn't blow before it damaged internal components. I live in Saudi Arabia so don't have a good repair shop here. The damaged components are diode bridges D301 and D302 and diodes D303 and D304. If someone has a parts list and ideally schematic I can access via PDF or via other internet means (or knows where I can find such information on line) I would be most grateful as I could then source bits for a repair. The parts don't have identifiable markings on them.
The instrument's owner will be coming back in 4 days and I'll be one dead donkey if it's not working.
Hoping someone can help ...EEE AWWW!:dunno:
 
Administrator
Joined 2004
Paid Member
Unfortunately it is quite probable that there is collateral damage, it is unlikely that the diodes failed due to excessive voltage, (unless HP used very low voltage diodes) more likely that they died from excessive current due to damaged/shorted electrolytics and other parts in the power supplies.. If you are very lucky it might just be the diodes and input filter caps.

Once you do the repairs I would use a series ballast lamp on the mains, 60W or so to limit further damage in the event you did not get all of the problems the first time around.

IIRC Every 339A I have ever seen has a voltage selector switch on the back panel so that unit should probably be switched for 220V operation. (or the right voltage for Saudi Arabia)

Good luck!
 
Last edited:
Reply to damaged HP339A posts

:)Thank you for your replies to my post. I have downloaded the service manual and have smuggled the 339A home where I can replace components. I have a short across CR301 diode bridge output - could be the bridge, CR303 56V zener, 2 capacitors or U301 LM325H voltage regulator. Needless to say I couldn't get any of these except the capacitors in the electronic shops downtown. Mail order will take a longer time, obviously. I will be unsoldering bits tomorrow to try and isolate the short and just pray the transformer isn't damaged. (The circuit diagram doesn't show secondary voltages at the transformer so I don't know what DC to expect at the diode bridges CR300 and CR301 outputs). Trying to get a replacement transformer quickly would be like trying to find a UK MP who hasn't fiddled expenses. Will post again if I manage to fix it.

Steve Cooke 22 April 2010
 
HP339A repair

After connecting the unit to 220V when it was set to 120V at the rear I found 56V zener diode CR303 was short circuit, happened before the rear fuse blew. This diode protects the Power supply IC for the meter circuit.
Naturally I can't get one in Saudi off the shelf so am waiting for one from the UK. As the diode is for protection against overvoltage only I just removed it for now. The unit now works fine but has less protection against overvoltage until I receive the diode. A lesson learnt!
Thanks again to those who replied to my original post and a big thank you to Agilent for making their old test equipment diagrams available. :D
 
Hi Steve,

I'm glad the damage was not serious and hopefully by now your 56V zener arrived.

I'm interested in DIY audio and after some research concluded that a good distortion analyzer with a sine wave oscillator will be very useful testing or building power amps. Initially I was looking at hp 239a which has impressive specs but unfortunately it doesn't have distortion measurement, hence my interest in finding a 339a - they're quiet expensive. I found one unit except most of the labels on the dials have worn off. I thought this was odd, I'm wondering if some solvent was used to loosen up these controls and it wiped out the scales. The unit apparently powers up but the seller claims nothing about testing and doesn't accept returns - typical fleaBay sales tactics.

May I ask what your application is where 339a is useful ?

Cheers
 
Hello Steve,

I got one HP339A and wonder whether is it normal for the instrument to behave like this on powering up.

When I turn it on, no matter what the input scale is, the meter swing to the very far end, then I turn it off immediately because I don't want to damage it. I wonder I should let its power on and let it have a chance to settle down.

Does your HP339A behave like this on powering up?

Thanks in advance.

Cheers,
WK
 
Administrator
Joined 2004
Paid Member
Hello Steve,

I got one HP339A and wonder whether is it normal for the instrument to behave like this on powering up.

When I turn it on, no matter what the input scale is, the meter swing to the very far end, then I turn it off immediately because I don't want to damage it. I wonder I should let its power on and let it have a chance to settle down.

Does your HP339A behave like this on powering up?

Thanks in advance.

Cheers,
WK


My recollection is that it does - I used one for the entire decade of the 1990s, but not since, so YMMV..
 
This is normal for 339A.

Looking at the meter amp I see why.

Until the Q1 Jfet is biased off the meter is at full deflection.

David.
 

Attachments

  • 339A meter amp.PNG
    339A meter amp.PNG
    25.6 KB · Views: 1,440
The mechanical design of the 339 and similar era gear was not HP's finest moment. The knobs are terrible -- weak, prone to break or crack, the dial skirts pop off the knob, and the printing is hopelessly fragile. I've used a very fine-point Sharpie to ink in the markings but it's a lousy fix.

On the good side, the 339 is a pretty good analyzer if you can live with about 0.001% as best resolution. The null floor is decent at about -110 to -115dB, so higher resolution is possible. The quad op-amp in the filter circuits needs to be replaced with a quieter unit, especially if you use the analyzer output to drive a wave or spectrum analyzer.

The oscillator is quite good -- better than the analyzer, which it should be for trustworthy use. The 239 oscillator is almost identical but is better in terms of noise, especially since you can isolate chassis ground (or not) with a front-panel switch. Nulling of the oscillator's 2nd H can be improved by padding one or the other of the 2k input/feedback gate resistors for the AGC JFET -- improvement can range up to 10dB. Doesn't affect 3rd H or higher.

My two cents...
 

6L6

Moderator
Joined 2010
Paid Member
I realize this is a bit of a thread necro, but there isn't a good reason to start a new thread...

Can the 339A measure IM distortion by itself?

THD seems nice, but talking to some people and reading about distortions for audio, it seems that IMD is much more important than THD...

The 339a seems to be a nifty tool, I am considering getting one for the bench, as without one, or something similar, you are blind to seeing distortion. I'm not a tech, just a hobbyist, but it's cheap enough that it looks like it could be useful.

Any ideas or comments? Thanks
 
I realize this is a bit of a thread necro, but there isn't a good reason to start a new thread...

Can the 339A measure IM distortion by itself?

THD seems nice, but talking to some people and reading about distortions for audio, it seems that IMD is much more important than THD...

The 339a seems to be a nifty tool, I am considering getting one for the bench, as without one, or something similar, you are blind to seeing distortion. I'm not a tech, just a hobbyist, but it's cheap enough that it looks like it could be useful.

Any ideas or comments? Thanks

NO. Only one oscillator and the required filters are no supplied.

Nothing special is required for IMD in the way of ultra low disto oscillators.
Have a look at the QA400 from QuantAsylum. It can do IMD and with an external notch filter and low disto oscillator it can measure below harmonics -120dB.

The 339A analyzer is a boat anchor.

Cheers,
 
Administrator
Joined 2004
Paid Member
The 339A is a boat anchor only in the sense that it is a bulky bench top unit, its performance is still quite respectable.. It will do thd+n measurements and has a 400Hz HPF, 30kHz and 80kHz LPFs as standard..

I used one all through the 1990s and were it not for the computer based FFT setup I generally use and a refurbished Amber 3501A I would definitely have one.

It will measure distortions levels reliably below 0.01% and has one of the cleanest and best performing sine wave oscillators I have ever seen. Residuals on the unit I used were around 0.002% or better which for the era was pretty good.

No IMD option however.

I would recommend a windows computer with Audiotester V3 and Pete Milletts Soundcard interface..

See my attachment for what is possible with a sound card. This is a capture of an acquisition loopback at 500mV and 10kHz.. Sample rate is 96kHz, and I can measure to a little over 40kHz, I can go faster if I need to do broader bandwidth measurements, but at some penalty in sfdr..

I have the luxury of dedicating an entire machine to this purpose although I will probably leverage it a bit for use with the uTracer tube tracing hardware I just bought.

This set up can also do multi-tone and any pair of frequencies within the allowable range based on sample rate. Below 10kHz and 2kHz at 1:1 just as an example. You can see sum and difference tones at 8kHz and 12kHz which will give you a sense of the IMD performance of the measurement setup.. Considering what it all cost it's quite decent. I'm still working on it.. :D
 

Attachments

  • screen capture 500mV.PNG
    screen capture 500mV.PNG
    126.1 KB · Views: 1,165
  • mutli-tone.png
    mutli-tone.png
    182 KB · Views: 1,139
Just a comment about IMD and THD. The causes of both forms of distortion are the same -- the non-linearities of amplifiers and passive components. In essence they measure the same things by different means. Neither form of measurement is preferable to the other, and both are limited in accuracy by noise and the performance of the analyzers used.

The HP 339 is not state-of-the-art, and hasn't been for a good while, but it is useful, especially for testing tube gear and much solid-state gear as well. The oscillator section is extremely good at frequencies below 10kHz. but if you intend to buy one, be prepared for some serious cleaning, repairing, and refurbishing -- most of the ones on the used market are not in good condition.