Audio Generator or Function Generator for amp testing & troubleshooting?

[MarkF786]

I've recently been getting into building and fixing tube amps, though I have some background in electronics. I've read in many places that for guitar tube amp maintenance, when using an o-scope an audio generator is preferable than a function generator due to the lower distortion on the audio generator, but other people say that with tube amps it doesn't matter either way because their THD is so high anyway.

Is there a benefit in using an audio generator, say for example with 0.02% distortion over a function generator with 1% distortion? Part of the reason I ask is because there are some pretty cheap DDS function generators that would save me some money instead of buying a half-decent audio generator. But at the same time, if an audio generator would be better in the long run, then it's worth spending the extra money.

Thoughts?

Thanks for the feedback.

Mark

[hasserl]

There are plenty of low cost audio signal generators available: audio signal generator | eBay

Or if you want to go real cheap, here's a freebie you can download to your Audio Frequency Generator FREEWARE

Or, download an app for your phone: App Store - AudioSigGen

[olddawg]

If you have an iPhone or a Droid phone or whatever, there are signal/function generator aps for free. You just plug the headphone out into the amp input and set the frequency, waveform, and the level with the phone controls. If you want to spend $1.99 to $15 or so they have some more professional looking aps that do the same thing. The are also free and inexpensive usable aps for spectrum analyzers and decibel meters. They are very handy to have in your pocket. I'm waiting to see a phone that has a built in scope and multimeter.

[MarkF786]

So there's no benefit in using dedicated audio generator over a digital function generator (including a PC or iPhone amp)? From what I've read, the audio generator gives you a larger voltage range and also will give you a cleaner waveform over a larger frequency response.

For a newb it's a bit confusing cause I'll see one person recommend getting an audio generator (and not a function generator or a PC/iPhone app) and someone else will say the opposite.

Heck, I don't mind spending $100-$200 to get an audio generator if it's the best long-term solution. Or if a $10 iPhone app will cover my needs, I'd obviously save money.

[tedmich]

I use these from Phil Marchand
Marchand Electronics: Sound card Function Generator, Real Time Analyzer, Audio Oscillator, Calculator, free download
not sure where they come from...

PS I got a mess of test tones on my iPhone but the horrid "20-20k sweep" and "Out of Phase Pink Noise" made it onto my "Metal" playlist.. gotta fix that soon!

[olddawg]

So there's no benefit in using dedicated audio generator over a digital function generator (including a PC or iPhone amp)? From what I've read, the audio generator gives you a larger voltage range and also will give you a cleaner waveform over a larger frequency response.

For a newb it's a bit confusing cause I'll see one person recommend getting an audio generator (and not a function generator or a PC/iPhone app) and someone else will say the opposite.

Heck, I don't mind spending $100-$200 to get an audio generator if it's the best long-term solution. Or if a $10 iPhone app will cover my needs, I'd obviously save money.

The iPhone Ap I use is called FreqGen. It's free and perfectly adequate for most audio applications. I do have a big old HP tube signal generator on my bench that I have had for years. Yes, it will put out a huge signal, I can sweep horn driver voice coils with it. But how often do you need to do that? And ....you can always amplify the signal from the iPod. The old HP isn't exactly portable either and is huge. But it looks impressive with it's big disc venier and leather handle.

[tedmich]

old freq generators are cool! My surplus shop (R5D3) gets tube units in on a regular basis and pulls the Telefunkin 12AX7s which Bob (bless his heart) sells for his std. $4

[Enzo]

There are no absolutes. WHat are you going top use it for? The guitar signal is typically under half a volt, maybe a whole volt if you bang on it. 1v of audio is what we consider line level. SO maybe some generator has a larger voltage range. So? If you never need more than a volt, what does it matter if one thing goes up tp only 2v and another goes up to 20v?

Avoid the temptation to take some individual specification out of context and decide it matters a lot. Dedicated audio - meaning sine wave - generators can be made to extremely low distortion specs. And for lab work, that is great. Function generators are a lot more versatile, but head to head maybe their "fleet" average is less super than the sine wave machines. But those things overlap. You $5000 function generator will run rings around your $80 sine wave generator. Wghen we are comparing fractions of a percent distortion, you will never hear the difference. If you want to design lab quality amplifiers or analyze super hifi circuits, then by all means get a super low distortion signal generator. But really, how pure is the sound from your guitar? That extra 1/8% distortion is not going to make a difference either way.

Most anything that spits out a reasonably clean sine wave will be useful.

In my shop I usually reach for a feed from the shop stereo receiver. I play music through the amps. I like it, music is a full range signal, and by listening I can hear the frequency response and any distortions. I have a sine wave available from a number of sources on my bench if I want one. I don;t know where it went when I moved the shop, but I used to have a small Yamaha personal keyboard, the kind Casio also makes. Small keys, power adaptor. It had a line out jack, and I often used that as a signal. I could get percussive sounds like piano, or steady sounds like flute or organ. If I wanted a steady tone, I could jam a little wad of paper between keys to hold one down. it even had a DEMO feature that played a tune if I wanted changing signal. A couple dollars at a yard sale, but quite useful.

Don;t overthink it.

[trevorus]

I have used an iPod signal generator, and one on my Android phone. The thing I like about them is they are stupid simple and CHEAP. You already have the device (in my case), and as Enzo said, THD is really a moot point on these sorts of devices when you are working on guitar amps. I use it to see where signal is dropping, or getting screwed up. So I can track a predictable sine wave from the input to output and see exactly what is happening to it. It's a diagnostic tool, not really a test device in the way of tuning hi-fi stuff. If you are looking for a super clean function generator, then you need proper equipment on the other end to listen to it for you, if you are getting that much precision.

[Steve Conner]

Soundcard/IPod generators are OK for casual work, but they have limitations.

They can't output a sine wave greater than half the sample rate, and as you get closer to that cutoff point, the wave can start to throb and beat in a weird way. Sample rate varies between 44.1 and 192kHz depending on how good a soundcard you have, so you can always get up to 20kHz, which is plenty for guitar amp servicing.

A possibly more significant issue is that they can't generate a clean square wave at all. It comes out all wobbly because of the Gibbs effect: File:Gibbs phenomenon 10.svg - Wikipedia, the free encyclopedia

So, an actual analog signal generator is a handy thing to have too. There are two kinds: RC oscillators that share DNA with the ancient HP things, and offer square and sine outputs, and "function generators" based on an IC that also generates a triangle wave. Both can be picked up on Ebay and at flea markets.

The RC oscillators can have as little as 0.01% THD if you're lucky: the function generators have 0.1 to 1% on the sine wave, because it's generated by warping the triangle wave.

Many of the RC oscillators have an annoying quirk: the level bounces up and down as you sweep the frequency. Function generators don't do that.

We work on lots of other stuff besides audio, so we have a selection of oscillators and signal generators that go all the way up to 1GHz. I also built my own super-low distortion audio oscillator (less than 0.002%) to go with a second-hand distortion analyzer that I picked up cheap.

It's impossible to make out small amounts of THD on a scope. I doubt you can see anything less than about 5%.

[nashvillebill]

I have a little function generator from Parts Express,the kit was something like $40. It does square, triangle, and sine waves. The sine wave isn't as clean as the sine wave from my old HP200CD sig gen but it's very usable (there's a tiny notch in the peak of the sine wave where the IC switches). The square wave was sharp and well defined on mine.

One advantage of the little cheap dedicated function generators, to me at least, is that if I do something really stupid and blow it up or drop it or melt it with the soldering iron, well I'm not out a lot of money. Whereas my smart phone would cost a fair bit to replace... I'm not so sure I would want to hook my smart phone up to a tube amp that has 500 volts on the B+, even with the best of precautions, because the way my life has been going lately, s**t happens and everything else in my life has gone to h*ll.

[km6xz]

I am sort of a test equipment junkie and like versatile gear for the bench. A lot of features are available with DDS but I find that overall, the best signal signal generator for a repair bench is also cheap. The features desired include low distortion(0.01% thd), accurate output metering in DBm and dbv, accurate stable frequency selection, and high output level and abilty to drive lower Z loads. I have lots of lab type generators which excel in one or more of each of those specs but the one that is used most is an old Heath IG5218. It can be found at Ham swap meets for $25-50 or $75-100 on eBay. In my old shop we had dozens of them and none ever failed. If one would have, all the parts are easy to find. It seems it was designed by people who actually knew what tasks a bench tech needed. With open or 600 internal termination, the 1% meter output accuracy is enough not have to have a True RMS meter hooked up across the output to see what you are sending. The frequency accuracy and resolution is quite good, using decade switches and one variable control of the 0-0.1 least significant digit. On this trip to and from the US, I am bringing another Sound Technology 1700B ultra low distortion generator and distortion analyzer to join the one I already have on my bench but I will still use the Heath more often for general repair and diagnosis because it saves time by giving all the information needed directly from the front panel which no other instrument, I have seen, does.

[Mark Black]

Some thoughts based on my own experiences...

In addition to the great info provided in these posts I would add, when looking to purchase:

Some signal & function generators can't be adjusted to low-enough output levels for high-gain work, so look at the range on both ends. Sure you can pad the signal, but that is an extra step and will change the circuit impedance. I also find it just a bit annoying (but maybe that's just me...).

Look for the presence of a secondary fixed-level output you can use to feed the trigger input of a scope. This will let you view waveforms in a much more stable manner. (I hate having the scope display jump all around until a signal threshold is met - maybe I'm just easily annoyed in my old age...)

[km6xz]

Yes, Mark, good to mention the attentuator output for triggering. The Heath I was describing has a separate square wave output with its own step and variable attenuator, perfect for triggering and not loading the sinewave signal path. The 10 db per step attenuator for the sinewave output goes from 50dbm to +20 with a variable for another 20db attenuation, all directly readable from the meter for a range of -70 to +23 into 600 ohms.
So a mic pre or other sensitive input driven by a low level low noise source or drive earphone or speakers at a low but useful level, directly. I have or had had just about all the HP, Wavetek, GR and Sound Tech generators and all require more steps to perform standard diagnostic measurements.

[MarkF786]

Thanks for all the feedback. I decided to first try some audio generators for my iPhone (including the Faber SignalSuite for $10), since many people recommended using one. One annoying thing I found is that the frequency of the signal is not rock solid; instead it vacillates slightly. Maybe this is due to the digital approximation of a given waveform.

When viewed with my DSO (Rigol DS1052E), the waveform moves around the screen giving the appearance that the phase occasionally flip-flops and the frequency counter jumps around; maybe this wouldn't be as much of a problem on an analog scope. When viewing a solid, steady waveform, my DSO stays perfectly locked on and the frequency doesn't jump, so it's not a problem with my scope.

And similar to what Mark Black mentioned above, when I reduce the output on the iPhone, the scope jumps all around because it can't get a good lock on the signal. It would be nice to have a trigger output to feed to the scope to remedy this issue.

It leaves me still wanting an analog audio generator. I've been looking at the Instek GAG-810 ($198) and the REK RAG-101 ($90). - but now that I look, they both only have a synchronization input, not an output to feed to a trigger. Maybe I should look to get a used Heathkit IG5218. Too bad I can't find anything new (and reasonably priced) that fits the bill since I feel like it's a gamble when buying old gear.