Attenuator for amp testing.

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Diabolical Artificer

Active member
Joined
Aug 12, 2014
Messages
32
When testing amplifiers etc sometimes we're dealing with outputs of between a few volts and 50v or more in some cases. Obviously we don't want to bung that straight into a bit of test gear like a spectrum analyser or audio interface. So how does one go about attenuating the OP without changing the load as seen by the amp and optimising the attenuator OP Z?

I've done a search online and here but can't find any reference to a DIY attenuator. Any search is festooned with guitar pedals, posh volume controls etc.

I have been using my Marconi TF2331 distortion meter's OP, but the input to this is 30v max.

Any thought's idea's most welcome.

Andy.
 
You can load the amp with a speaker, and a resistive pad. The pad can be high enough impedance to not add extra load on the amp, while low enough to feed the test equipment.

Note: you may want to put the speaker in a closet or other room as it can get loud.

JR
 
I've posted this project elsewhere on the forum but something like this is the ideal solution.

https://linearaudio.nl/la-autoranger

On my dScope or the AP I owned previously the input attenuation is automatic (and compensated for in software).  Jan Didden's project is an effort to make that functionality available to everyone using typical sound card + software testing for home.

He's gearing up to launch soon.
 
You might find one of the larger Daven power meters fairly cheap.  There's one up to 100W.  Combination of resistive and inductive loading.  They let you set output Z.  A bridging pad off of that would be a nice combo. 
 
Just put a large power resistor across the load and then add a conventional log potentiometer across that to tap into the signal.

Although in practice it's usually more efficient to put 2-3 power resistors inside a metal box along with some fins in which case you might as well add the potentiometer across only 1 of the resistors (probably the last one connected to ground just in case ground is actually earth ground like in a Fender with a Presence control). I would isolate the "ground" from the metal box though (meaning use plastic isolated jacks).

Not that a conventional pot would not be low enough impedance to drive a speaker. But a 1K pot would give you a source Z of at most 250 ohms which is just fine for most testing. If you want constant source impedance you can do that with a slightly fancier circuit.

I made rotary switch based attenuator with all really low Z resistors so that it could drive a speaker but I made a mistake in that when switched the load could be removed momentarily from the amp. One time I switched it with the amp at full power with a test signal. The EL34s flashed several times and died miserably. It was very exciting.
 
Thanks for your replies. I'd be using a dummy load John most of the time. I have four 100w 2r resistors mounted on a forced air cooled heatsink off a welder. I know what you mean about testing with a speaker, I use ear plugs.

The LA Autoranger looks nice and just the job, but being perennially skint, it's out of the question. shame, it looks just the job.

I have a pretty good junk box with several big WW R's and various 5W 10 turn pots, so will probably go with a homemade resistive pad and maybe add a couple of transformers for 600 ohm OP. Will keep you all updated.

Cheers, Andy.
 
DrWobble said:
When testing amplifiers etc sometimes we're dealing with outputs of between a few volts and 50v or more in some cases. Obviously we don't want to bung that straight into a bit of test gear like a spectrum analyser or audio interface. So how does one go about attenuating the OP without changing the load as seen by the amp and optimising the attenuator OP Z?

I've done a search online and here but can't find any reference to a DIY attenuator. Any search is festooned with guitar pedals, posh volume controls etc.

I have been using my Marconi TF2331 distortion meter's OP, but the input to this is 30v max.

Any thought's idea's most welcome.

Andy.
It's traditionally done with a set of power resistors that can be connected in parallels for adjusting the load value.
I use 4x 600W 8.2 ohms resistors that a I can use for loading two channels into 4 ohms or one channel into 2 ohms. That's because I test very powerful class-D amps; if you tests valve amps, you may use resistors with a lesser power rating.
I have a fixed 40 dB attenuator using a 100k in series and a 1.1k with a 22k trimmer across, so it can be adjusted exactly. The resulting 1k source impedance ensures the input capacitance of the audio analyser does not interfere significantly.
Why 40dB and not 20, for example? Because with a 10k/1k attenuator and 100V rms, the dissipation in the 10k resistor would be 1W, which could introduce temperature-related errors, even if the resistor was well-dimensioned. With 100k, dissipation is 0.1W.
 
I  made a load-box with a pair of 8ohm, 50W aluminium ceramic resistors mounted in a decent sized aluminium box.

Each of the 8ohm/50W resistors are brought out to banana plug red, black terminals on the top.

I can then use short banana plug leads to make combinations of 4ohm/100W (parallel) or 16ohm/100W (series) or stereo 8ohm/50W  loads. I connect these terminals to the 'speaker output' of an amp-under-test.

Then I add across each power resistor, 3 resistors in series, with the middle one set for 10Kohms 1W and a 'balanced' XLR connector hot (+) and cold (-) across it - that feeds my audio interface input.

The other two resistors hang off each end of the 10Kohms resistor back to the load - you can use whatever attenuation ratio you want to suit your audio interface  input range.

In my case, I like the load box to attenuate say 50Vpp down to the audio interface input max of 10Vpp or so.

I also have a BNC connection for my cro and/or voltmeter off the load box 'raw input' terminals.

ON top of that, I have a 'high voltage' measurement box, with a safely insulated 'probe' and 'ground' cable (a decent aftermarket cro probe) into an aluminium box.

The probe BNC goes inside the alu box in series with a 630Vdc 10uF cap and a 2M 1W resistor to ground - that gives a safe, hiZ ac-coupled measurement point.

I add the same kind of thing as the load box - 3 resistors in series across the measurement point to ground, with the middle resistor being 10K (which the audio interface likes).

The other two resistors, off the 10K are chosen for the attenuation ratio desired - I like to measure up to 450Vac or so and map down to the audio interface 10Vpp range.

Because the HV measurement range depends a lot on the circuit-under-test, I use a 2 pole rotary switch to give me 6 positions corresponding to divide by 45  down to divide by 2.

Again, I have a bnc connector to allow a parallel CRO/voltmeter feed of the raw input.


:)  Works well for me in most cases where the 10uF+2M2 input Z doesn't marr the measurement of the circuit under test too much.

PS - one needs to remember to discharge the HV measure probe to ground after each measurement is completed  :)

PPS - this method is pretty safe for the audio interface inputs, but I recommend a reasonably cheap audio interface just in case.

I am  now using an M-Audio Profire 610  interface - it's pretty cheap, has balanced I/O and good performance plus solid drivers for my Win 7_64 AMD box with cheapo firewire card.

Using this setup and with awesome software app 'Room Eq Wizard REW',  I can do THD measurements of amps and tube stages with pretty great accuracy on the super cheap.

You can improve the accuracy by using finer tolerance load resistors (expensive) or by paralleling regular tolerance resistors.
 
Sorry for just seeing this, but if you think the Autoranger is just the ticket but too expensive, you might look at Pete Milliett's sound card interface. Looks like much the same thing at quick glance. I built one for about $80 all in. You could go cheaper if you use an existing enclosure and have some of the parts in your pantry already. If you had to buy all the parts and didn't leave off extras like the panel meter, it would be closer to $170 by his estimation.  Still cheaper than the Autoranger. Just a thought. 

http://www.pmillett.com/ATEST.htm

BT
 
ruairioflaherty said:
On my dScope or the AP I owned previously the input attenuation is automatic (and compensated for in software).  Jan Didden's project is an effort to make that functionality available to everyone using typical sound card + software testing for home.
An autoranger is a nice commodity for those who make repeated measurements and want to improve their workflow. It does not make much sense when your time is almost free.
The suggestion to use a properly dimensioned load resistor (or combination of) and tapping the signal via a pad is a good solution for a DIYer; however, depending on the power of the amps to be tested, a potentiometer may not be adequate. A 250W/8 ohms amp delivers 44Vrms; if inadvertently put into clipping, the square-wave power doubles. With a 1k potentiometer, that's 2W and 4W respectively. Most potentiometers will fry. I think it is much better using fixed power resistors (wire-wound ceramic are quite inexpensive).
 
I've been reworking my attenuator box which I use to allow me to measure large ac signals with my M-Audio Profire 610 firewire interface and Room EQ Wizard ....

I'm interested in measuring noise floor for hum (low freqs), hiss (mid to high freqs) as well as THD (distortion components and total) and freq response (sweeps).

This is for tube circuits, typically with high impedance test points where there are high dc voltages present as well as ac signals of high amplitude.

Plate circuits and the like, with ac and dc of hundreds of volts typically.

:)

Recapping previous discussions, an attenuator is needed to 'map' large ac signals to the available input voltage range of the audio interface. 

The max input signal REW allows is -3 dBFS. The input is balanced and so that corresponds to around 13.4Vpp differential voltage applied, for the M-Audio Profire 610 'line' input.

The 'line' inputs don't have a gain adjust, so it is fixed - this makes things simpler in some respects.

SO - I want an attenuator box with :

- cheap CRO probe with bnc connector as input
- big cap for blocking of any dc that may be at the circuit measurement point
- xlr balanced output connector
- ground 'lift' switch for the output xlr ground pin 1 to test box chassis

and

- a switch for varying the attenuation
- a high impedance at the probe so as not to disturb the signal under test as least as practicable
- a low impedance at the attenuator box output to allow feeding the line input without losing bass

These last 3 conditions are a trade off against each other, so the trick is to find a reasonable compromise.

So far I'm using a simple set up like this ...


 

Attachments

  • atten_box_pic1.jpg
    atten_box_pic1.jpg
    96 KB · Views: 19
So the output voltage is equal to the input voltage divided by the ratio 2X / (2X + 2Y)

Now the CRO probe typically has an impedance of 1Mohm.

(some have a x10 switch which alters that to 10Mohms)

So the impedance presented to the test point is 1M // total impedance of the attenuator.

For testing plate circuits, of high-ish impedance,  typically hundreds Kohms, we need the probe impedance to be as high as possible, so as not to overly load the circuit-under-test, which would reduce it's voltage signal as well as increase it's distortion.

That means *high* Y resistors.

However, we want the impedance fed to the audio interface balanced input  to relatively low, in the tens K, otherwise we begin to run into impedance mismatch issues, which upsets our freq response sweeps.

The M-Audio Profire  balanced 'line inputs' has a spec of ' > 20Kohms', so a sensible maximum source impedance would be 20Kohms.

That means X resistors of 10K each, for a total of 20K 'balanced'.

[ the junction of the X resistors is connected to XLR ground ... which is the M-Audio interface's ground ]
[one can have a 'ground lift' switch which makes/breaks connection of this junction to the box chassis]


OK then, the problem is that in order to maintain the  X = 10Kohms and Y= *high* Kohms conditions, our attenuation ratio becomes quite high.

It gets difficult to achieve the lower attenuation ratios without resulting in a low probe impedance which then excessively loads the circuit under test.

That makes, for lower amplitude signals, say 50Vpp max,  achieving a decent 'signal to noise ratio' into REW an issue - noise can begin to dominate proceedings as well as distortion increasing  because of A/D resolution issues.

The higher attenuation ratios, say for signals of 100Vpp and greater is not too difficult and can be achieved without undue loading effects.


:)

In my box, I have a rotary switch which varies the two Y resistors and the X resistors are constant.
Varying the X resistors (impedance seen by interface input) complicates things some more :)

I have six available settings - so far I have used

i)    Y = 51K    , X= 10K ;  Idealised atten ratio div 6  ;      Probe impedance approx 1M // 122K  -> 108K
ii)  Y = 180K,  X=10K ;  Idealised atten ratio div 10  ;  Probe impedance approx 1M // 380K  -> 275K
ii)  Y = 470K,  X=10K ;  Idealised atten ratio div 25  ;  Probe impedance approx 1M // 960K  -> 490K

In practice, for plate circuits, setting i) has around 24% loading effect and setting iii) around 6%

It's a bit approximate in practice - but I aim for the highest signal I can get into REW, around -20dbFS as a reasonable minimum  without loading the circuit under test more than around  10%.

Checking freq response with sweeps shows no undue rolloffs 20-20KHz bandwidths and frequently 10 .. 44KHz BW in typical plate circuits.

...

One can use more elaborate attenuation schemes, perhaps with both X and Y being variable on the switch to improve things  ..  it does get more complicted fairly quickly  :)
 

Attachments

  • atten_box_pic2.jpg
    atten_box_pic2.jpg
    156.7 KB · Views: 5
As I said, it's a little approximate, with the REW signal amplitude and THD displays affected by both the circuit-under-test loading as well as the approximate attenuation factor applied....

THD measurements are very amplitude/impedance dependent at the best of times ...

but with a little 'standardisation', reasonably consistent results can be obtained with enough 'relative' validity to provide valuable assistance in fine tuning tube circuits and such.

One quickly gets a sense of 'benchmark thd' measures for determining when tube circuits are 'happy' or not.

In the past I used a single setting - now I'm looking a bit more closely at the results and how valid they are :)

....

Similar principle is used for a 'load box' for testing of amplifier outputs ...  with suitable high power non-inductive resistors and impedances set around common speaker loads.

...

It would be nice to have a proper THD measuring  setup like the pro's  :)

....

Also of interest would be some DIY test setup  - maybe with mosfet followers as impedance converters and variable, low distrtion gain to optimise singal-noise ratio into REW  ....


PS - although the box has a 'bleeder resistor' effect, one needs to remember to *discharge the cap* after taking measurements by touching the CRO probe tip to ground.

That big cap can charge up to high voltages! that can take 10s or so to 'bleed away' naturally.

...
Please feel free to comment  ....
 
The 'load box' for amp testing goes like this ...

I use a metal box with  two seperate  pairs of black and red 'speaker' type terminals.
Across each of the pairs is an 8ohm 50W ceramic power resistors.  (I'm too lax to get non-inductive types)

I can use the box for stereo 8ohm 50W loads or I can jumper them with wires to give mono 4ohms 100W or mono 16ohms 100W.

Across each of the speaker terminal pairs (with the 8ohm 50W resistor ), I also have a phono type unbalanced connector to feed a paralled speaker, for monitoring purposes.

As well I have an XLR connector to feed the audio interface 'line input'.

To feed that balanced XLR connector, I put 0.6W resistors across the 8ohm  ...  they are the 'X' and 'Y' resistors as I described before.

This time, one needs to have the total attenuator impedance to be large, compared to the 8ohm power resistor.

Again, X = 10K for a total of 20K 'balanced' impedance feeding the audio interface 'line inputs'.

Y resistors are set for the desired attenuation ratio - there is no problem with the impedance presented to the amp-under-test; it sees 8ohms // 2X + 2Y      where    2X+2Y  is a lot higher than 8ohms, so it has negligible loading effect on the amp-under-test.

I have a single attenuation ratio set, good for about 150Wrms into 8ohms; that equates to 34Vrms or approx 103Vpp
.
So I need to 'map'  103Vpp  to my audio interface max range of 13.4Vpp (balanced) - an attenuation ratio of approx div 7.5.

That winds up as Y=68K and X=10K for a total attenuator impedance of 8ohms // 156K -> 7.999ohms  :)
Attenuation ratio is approx 20K/156K -> div 7.8

Again. one can have a rotary switch, varying the Y resistors and holding X resistors constant - this allows more fine tuning of amp power testing, say 50W, 100W, 150W into 8ohms.

And that's the whys and wherefores of  how to make a balanced HV test box and a balanced amp dummy load box.

One can adapt what I've described to provide for testing of opamp and solid state circuits, which typically have a much lower impedance than tube circuits, meaning it's relatively easy to get attenuation ratios suitable for say 30Vpp under test  ie. div 2.5 or so.


:)

 
It's hard to understate the value of these most valuable test tools pursuant of fine tuning of one's circuits and amps  ...  :)

THD measures can be quite enlightening, to say the least - one can tweak things with deep visibility, to quite an impressive state of performance.

Before the advent of the most awesome 'Room EQ Wizard' app's THD displays, I was mostly blind  8)

With REW and the test boxes, I can see .. and therefore 'hear'  what's what!   

That leads to the 'Long and Winding Road' of linking performance measures to 'good sound'  ;D

...

I leave it to the brighter lights of this esteemed GroupDIY to discuss the finer points of all this.

There are, of course, good reasons why pro test rigs cost a bundle  and are used by professionals  :)
 
And a quick final note ...  I ended making an 'op-amp' test rig to measure 'how low is low'  with respect to THD ...

I made a 1ru box with a JLM psu board for +/- 18Vdc rails and added some cheap eb*y modules for the functions of balancing/unbalancing, inverting/non-inverting buffering and variable gain.

That allowed me to get a sense of the lower limit of THD and how it relates to the audio interface's 'residual  THD and noise'.

It gave good enough results to discriminate between regular audio opamps like the venerable ne553x and tl07x series and the uber opamps of todays world.

Also interesting, is to quantify  the performance of modern line transceiver chips like those of That.

Really interesting and amazing as to the level of performance one can achieve, even 'on the cheap'

At That end of town, a high performance audio interface is needed ... like the pricier offerings from RME, Motu and the more modern Focusrite  etc .... and beyond into the price stratosphere  ???
 
alexc said:
And a quick final note ...  I ended making an 'op-amp' test rig to measure 'how low is low'  with respect to THD ...

I made a 1ru box with a JLM psu board for +/- 18Vdc rails and added some cheap eb*y modules for the functions of balancing/unbalancing, inverting/non-inverting buffering and variable gain.

That allowed me to get a sense of the lower limit of THD and how it relates to the audio interface's 'residual  THD and noise'.

It gave good enough results to discriminate between regular audio opamps like the venerable ne553x and tl07x series and the uber opamps of todays world.
So did I read this right? Are you're running these op-amps off of +/- 18V? Or is that voltage going to some unmentioned 15V regulators that then power the op-amps?

The 553x series have a max voltage of +/-22V, but not the TL07x, where +/-18V is the "Absolute Maximum Rating."

I'm thinking of this thread:
https://groupdiy.com/index.php?topic=63399.0
 
I run the ne553x and tl07x series off +/-17V generally. 

I use the higher rails on the more modern opamps, such as the lme49990 which can go up to +/-22V.

The psu I use has adjustable rails.
 
Back
Top