Troubleshooting Audio Amp

Does it become good and bad at the same places each time you play the same tape? If it does, the problen is in the recording.

Make a probe to feed the signal into on a high-impedance, high-gain amplifier and trace the signal through the machine. Is it diistorted at the playback head? At the volume control? At the line-level output? At the loudspeaker terminals (you could have a damaged loudspeaker)?

Yes it does. But that recording still had to be processed through the amp chain.

A scope would function in such a role just the same AFAICS., would it not? I need to make up a test tape with a constant high quality tone. I can do that with my Ferrograph and audio sig gen. Trying to discern which sections are distorting using rock music for the purpose has not worked out too well - even though it wasn't Black Sabbath!

I can't follow your logic. If the tape sounds good in some places and distorted in others - and these are the the same places each time you play it - the fault is clearly something to do with the tape and not the amplifier.

Turn down the volume and listen carefully to the tape transport - is it squealing during the distorted passages? Does the tape each side of the head block appear to be running normally or is it shuddering because of patches of sticky-shed? Rest your finger gently on the feed spool and feel for any vibration, then try gently putting a bit of load on it until the tape slows down a fraction. If it is unduly sensitive to the load test or stops dead with little pressure, there is some sticky gunge lurking somewhere you haven't looked yet.

Depending on the type of distortion, it might be quite difficult to see on a 'scope but very obvious to hear through an amplifier. On a 'scope, distortion of 10% on a sinewave will show up if you know what you are looking for - but on complex programme material even 30% may not be detectable on a 'scope trace. On a listening test, as little as 2% is detectable on sinewaves and 'clean' programme material - but with material that has been distorted for 'musical' effect, all bets are off.

Makes no sense to me, that observation. The tape is brand new and from a well-respected manufacturer. Think about it, the signal on the tape didn't get 'impressed' there direct from a microphone or whatever, it had to be amplified. It had to pass through the amplifier chain and be mixed with the output from the bias oscillator, so whatever defects were present at that moment from moment to moment from the amp as the tape passed, will be faithfully recorded and re-playable. I assume we're not on the same page, so will not address your follow-up until we're straight on this one.

My apologies, I was assuming you were playng back a tape that had been recorded some time ago and it had recently begun sounding distorted. I hadn't realised you were looking for the cause of distortion in the

In that case you need to monitor the signal going to the recording head (preferably just before the point where the high-frequency bias is added). If that is clean, suspect there is poor contact between the head and the tape. The head-tape contact is far more critical on record than on playback. If the bias is varying, that can cause all sorts of problems but that is much less likely to be happening. Frequency modulation of the sound by tape head squeal is another possibility, especially with new tape that might have a manufacturing defect, or very old tape that has begun to deteriorate.

Dirty contacts in the record/playback switch can manifest themselves as variations in signal level and distortion and are more likley to occur in a machine that hasn't been used for recording for a long period.. ....But, the head-tape contact should be your first suspect.

My previous remarks about distortion being much easier to detect by listening, rather than with an oscilloscope, still apply. If you have a sinewave signal generator, that will make it much easier to hear distortion than trying to spot subtle changes by listening to programme material.

Okay; noted, many thanks.

Thinking about your very valid suggestion about using the sound rather than looking at a waveform, I'm guessing a crystal earpiece would work just as well rather than making up a hi-z probe+ amp would it not?

You really need something that gives good sound quality so that you can hear the difference when the distortion occurs. My experience of most high-sensitivity earpieces siggests that they could give more distortion on undistorted signals than the distortion you are listening for - so you may not hear much difference.

If your distortion is really gross and isn't likely to be caused by the loading of the amplifier by the earpiece itself, you might be able to hear it and track it down by that method. Although crystal earpieces may read open-circuit on a DC test, they have a fair bit of capacitance and long leads, which could upset the circuit you are trying to monitor.

You haven't really described the distortion but if it sounds like harsh bursts of sound on loud notes with almost silence in between, that is usually an indication of a weak or failed bias oscillator. Overloading the recording gives very sudden onset of bad distortion on loud peaks, so watch your recording level meter closely (assuming it is correctly calibrated).

Good point; I had kind of forgotten about the sound quality of those things over the years since I last used one (or should I say decades).

Yeah, I'm going to forget that approach altogether. Shame, but you're quite right of course.

No, it's not coming in bursts. As I said in my original post, it really does sound for all the world like an old 45 vinyl record that's been played with a worn stylus. Then a few minutes later it might be fine again - for a while. I made up a test tape of a 5khz sine wave using my Ferrograph deck and played it on the Uher. I could hear the distinct difference in quality cutting in and out periodically when I played it back. I then tried to make another test tape of the same tone on the Uher to play back on the F'graph, but didn't have sufficiently low source impedance from the generator in this instance so I had to use a radio broadcast instead. Not ideal, but the distortion is still readily noticeable on high quality audio recordings. Unfortunately, however, even on the lowest speed of 3.5ips (IIRC) to was still whizzing past way too fast to be of any use. I'd recorded it using the second lowest of the Uher's four speeds and there's no point trying again on the lowest

That distortion is a combination of intermodulation and noise-behind-signal. Really severe tape squeal is the only thing I can think of that sounds like that on a tape recording

Have I understood this correctly?

2) A tape recorded on the Uher shows similar intermittent distortion when played back on the Ferrograph.

By elimination: 3) rules out any problems in either playback amplifier,

Years ago there were problems when one or two tape manufacturers produced tape with a backing that caused squeal in pressure-pad machines (which both the Uher and the Ferrograph are) but you would be able to hear that by turning down the sound and listening to the tape going past the heads.

I'm surprised the signal generator didn't have low enough source impedance to feed a signal into the Uher; most audio signal generators have an output impedance around 600 ohms and most Uhers have inputs of around 600 ohms and 'high' (around 47k).

Why were you using 5Kc/s as your test tone? The second harmonic of that, which is the first spurious frequency created by distortion, is

Thoroughly clean the pinchwheel. It could be accumalating microscopic bits of ferrite embedded in the rubber. Chech one of the tapes on another machine as if that is the case you will be corrupting all your tapes each time you play them as the ferrite is in a sense recording or deleting microscopically on each passage

Thank you for your observations. I had hoped to devote some time to this issue today, but other events kind of took over, so the next slot will be Weds afternoon. It's a real nuisance not being able to record the test tone on the Uher. I'm surprised as well. In fact I'm wondering if there's something else going on there. I'm going to follow up on your suggestion of cobbling together a hi-z amp for probing stage-by-stage. I think I've got some FET front-ended audio op amps somewhere in my parts stash which should be ideal for this purpose. Can't recall what I originally bought them for, but whatever it was never got finished, clearly! I did notice that for some reason on the slowest speed (the 15/16"ips) there's considerable variation in the tonal reproduction with music off the radio, as if the speed is wandering during transport. Not noticed any such effect at the higher speeds so not sure what that's all about. Fortunately, it's only music where it's noticeable; speech just sounds normal and at that speed, it's really only speech that one would use it for anyway.

Yes, that's a valid point. I don't think it's an issue in my case as I'm only using high quality tape which is not depositing anything visible on the pinch roller/capstan head. I suppose there's an outside chance something invisible to the naked eye could cause problems, so I'll give 'em another clean up just to be on the safe side. I had some Ampex tape and that was vile for sticky shed syndrome; the oxide layer came off in big old gummy shards. Fortunately, I'd been forewarned about Ampex tape, so unwound the reel by hand (ended up with sticky brown hands for my trouble!)

Have you got the correct connections to the input socket?

Easier than that: connect a piece of screened cable to the input of an amplifier with reasonable gain and make a probe from a 1 megohm resistor in series with a 10 nf capacitor at the free end. For an earth connection you can use a crocodile clip, either on the end of a long piece of wire coming from the amplifier or on a shorter piece connected to the screen of the cable.

Speed variations are much more obvious at slower tape speeds, 15/16 ips wouldn't normally be used for music. 3+3/4 ips is the bare minimum for domestic music and 7+1/2 ips is the minimum for professional recording. I recorded a few folk music items at 7+1/2 ips for local radio and they were happy enough with that - but for orchestral music nothing less than

The speed variation could be caused by dirt on the surface of the capstan or pinch wheel, or their bearings could be getting a bit tight (unlikely for the capstan). The pinch wheel or the tyre of the flywheel could be a bit dirty or oily - or it could have an indentation where the machine has been left 'in gear' with the motor not running for long period of storage. Other casues include dragging friction clutches, unevenly-wound spools and binding in the guides.

The frequency of the speed variations and their 'waveform' can often give a clue to their cause. Does the speed vary smoothly and sinusoidally (unven pinch wheel) or does it 'hiccup' slightly (dirt or indentations)? Watch the rotating components and try to work out which one's rotation corresponds to the speed of variation.

Yes, but there's an adaptor inline that could be the cause of it. I'll investigate that tomorrow. One of the really old Chinese things: prime suspect. Chinese stuff continues to improve, but god knows it was

Delightfully old school. I like it! What you're describing sounds rather like an oscilloscope probe - except with much higher capacitance. Shame as a scope probe is very ergonomic and ideal for this kind of poking around.

So you're saying the speed's varying at all the higher speeds as well, then? I guess that would make sense given the relatively agricultural mechanics of these old decks, but I still can't hear such a wobble at any of the 3 higher speeds. I suppose a frequency counter and test tape would be the answer to finding out more about that.

I haven't been able to tell - but I haven't tried the constant tone test tape yet. That will show it up much more clearly and that's what I intend to do tomorrow if I can nail the distortion issue, which is the more significant problem right now. Thanks for all the tips. You clearly know your stuff!

The capacitance I am suggesting is in series with the signal to stop any DC from getting into the test amplifier. The capacitance quoted for oscilloscopes and probes is the shunt capacitance from signal to earth. You could use an oscillocope probe if you have one with a suitable plug, but interpose a 10nf capacitor between the probe tip and the point in the circuit you are measuring, so as to block any DC.

If you want an ergonomic probe, put the components into a ball-point pen case (if you can find a small enough capacitor). ...or just put the resistor in the pen case and have the capacitor in the plug at the amplifier end of the cable.

I have a home-built test amplifier screwed to the underside of the shelf above the workshop bench, with a veriety of input connectors and probes. It has two attenuators calibrated in 10 and 1 dB steps, with a meter that reads directly in dB or can be switched with an electronic zero to read variations of 0.1 dB. There is an output at 0dBm level for feeding external equipment, such as high-quality headphones or an oscilloscope, and it has a small loudspeaker, which is often more use than the oscilloscope for identifying the cause of fault.

At higher speeds the flywheel effect of the rotating components is much greater and any given actual speed variation is a smaller percentage of the faster running speed - so both effect improve the overall speed stability.

Hiccups due to dirt on the capstan will be the same magnitude at all speeds but will occur at a faster rate at higher capstan and tape speeds..

Remember, the wow could be coming from the machine that recorded the tape, so try recording one on each machine and playing them both back on one machine - does the wow and flutter sound the same for both recordings or different? Then try the same test for one recording played back on both machines.

If you can remember what you are doing without getting muddled (a notepad might help) you may be able to draw some useful conclusions about which machine is causing which effect. Don't expect to make sense of this when you are tired or have other things on your mind.

You will be lucky to record 5Kc/s steadily at 15/16 ips. and you are unlikely to hear harmonic distortion from it at any tape speed. If you can hear something that sounds like distortion on a 5Kc/s tone, it is almost certainly not waveform distortion but more probably something else, like noise behind signal . If it contains lower frequency components than the initial tone it might be due to intermodulation, but that ought not to occur on a single sinewave. Try 630c/s and see if the results are different.

Long ago I was given a Harman Kardon 505 amp.

It occasionally seemed to "lock up" at full DC output on one channel.

With a colleague reading off test voltages from the schematic I probed the circuit, eventually getting to the tone controls. The solder joints all looked perfect, beautiful fillets, no dewets. The meter probe touched a lead (I think from a transistor) and the lead slipped in the fillet!

Apparently the plating had corroded or otherwise released the transistor lead. I reflowed the solder joint (and maybe a few others, don't remember) and the amp is still working 30 years later.

The surprise wasn't so much the faulty joint but the visual perfection of the fault. I'm not at all certain the story is relevant to the problem at hand, but if all else fails a "push test" on the solder joints in the record amp chain might be worth a try.

Thanks for reading,

bob prohaska

Thank you, Bob. I've not encountered that particular phenomenon before, but it does go to show how you can never make assumptions on connectivity from visual inspection alone. Thanks for posting as it's very worthwhile to remember this particular pitfall could be the culprit.

Oh dear. Here's the probe I made up today. Can you spot where I went wrong? Sigh.....