Hammond AO-29 Frequency Response

[stefanv]

I'm in the midst of a ground-up rebuild of a Hammond AO-29-13 amplifier, as used in a 1962 M-100 series. I actually bought a second AO-29 amplifier, and am using the chassis, most of the tube sockets, and the percussion transformers from that, taking only the power and output transformers from my original AO-29 once I'm ready.

As part of the rebuild, I'm doing extensive before and after measurements of the performance of the amplifier, so that (a) I can see the rebuilt one is working right, and (b) see if anything changed. I thought I'd share some of my findings here.

First, here is the end-to-end frequency response of the stock amplifier (electrolytics are new), with the pedal at full volume, and the Volume Soft switch up (i.e. volume not soft):



You can clearly see the 6dB/octave roll-off at 3.5kHz, beginning gradually at about 2kHz, that Hammond designed in to minimize key click.

More to follow in later posts ...

[stefanv]

Here is the end-to-end frequency response with the pedal at the lowest volume,



and here is the difference in frequency response between the two:



You can see that reducing the volume with the pedal changes the tone, de-emphasizing the mid-range.

All plots are normalized to a peak of 0dB. These are not absolute levels.

[Wes]

Signal generator or tonewheel generator input?

Very awesome graphs, btw!

[enor]

Nice work! I love it when science confirms what my ears tell me...

[stefanv]

Signal generator or tonewheel generator input?

White noise, piped into the "A" input terminal. I wanted to measure the characteristics of the amplifier, not the organ as a whole. The tone generator of course gets louder as you go up in frequency, to compensate for the amplifier's roll-off.

[stefanv]

Volume Soft

Here we have the response with the pedal loud, and the Volume Soft switch down (on),



and this is the relative effect that the Volume Soft switch has:

[Wes]

Excellent methodology! Nice to have you back aboard, Stefan.

Wes

[stefanv]

Vibrato

All the results above have been from the "A" input to the speakers (equipped with a line-out tap). I also did some testing from the "B" input, which feeds into a preamp much like the "A" input, but then passes through the vibrato system before entering the intermediate amplifier. The following thus include the "B" input preamp, the vibrato delay line, and the scanner.

Please keep in mind that my vibrato delay line is modified, so a stock M-100 will have slightly different results. I wish I'd had the foresight and equipment to do these tests before the modifications.









The last one is interesting. The Celeste settings feed some of the output from the terminated end of the delay line back into the input. Because of the delay, this means that some of the fed back signal will be exactly 180 degrees out of phase with the original, resulting in cancellation. The effect is a comb filter.

[geoffbrown]

What are you trying to achieve with your rebuild ? is it to give a more B3 like sound or something else?

[stefanv]

What are you trying to achieve with your rebuild ? is it to give a more B3 like sound or something else?

The old amp is very noisy (scratching and rumbling noises), I suspect due to old resistors. Rather than replace them one at a time, I just decided to bite the bullet and start from scratch.

[geoffbrown]

Ok I was assuming, going by your graphs that you were trying to change the characteristics of the amp.

A year ago the PT my M102 blew and due to my location, spare parts are non existent, so I used a substitute PT from a fender bassman guitar amp, this has resulted in vastly improved bass , a less muddy sound and much more pronounced percussion.
So I think there is considerable scope to experiment and make improvements to the original amp.
I'm currently building an adjustable voltage regulated power supply for my m102 ,like many guitar amps have.

[stefanv]

Geoff, the graphing is primarily so I can see I can compare the performance of the new amp with the old to ensure everything is working as expected. I am making one small change, which is to increase the high end response of the output amplifier. This will brighten up the organ a bit, and also make the signal path from the line-in to the output a bit more hi-fi, so that if I feed my synth input into it, it won't muddy up the synth sound.

- - - Updated - - -

Time for a few more response curves. These show the input (up to and including the volume pedal) and output (line-in to speaker out) stages of the amp separately:





Combined, they give the overall response that I posted in my original post.

[stefanv]

A Break from the Graphs

Earlier today, I completed the rebuild by installing the power and output transformers from the original amplifier into the rebuilt one. Here are a few photos. First is the underside, where you can see that I've done away with the turret boards, and wired everything point-to-point, using terminal strips:





One mod I made to the amplifier is to fuse everything. There are fuses in all three pairs of transformer output leads (320V, 5V, and 6V), as well as a fuse in the hot input lead:



Here's a top view of the completed amp, ready to be reinstalled. The multi-cap is new:



After I reassembled the amp, I bench tested it with an old hi-fi speaker, and Klaus Wunderlich piped into the "A" input:

[geoffbrown]

The extra fusing is a great idea ,something which I will do to my amp next time it's out,my m102 had a fuse in the reverb amp but nothing in the main amp,I'll be watching this thread with great interest

[stefanv]

The rebuilt amp works! Here's the end-to-end frequency response with pedal loud and volume loud, with the old one faintly overlaid in light blue:



You can see that the overall response is pretty similar below about 17kHz.

Here are the curves for the first and second stages alone:



The first stage seems to drop off a little more rapidy, and doesn't bounce back after 12kHz (which is beyond the range of frequencies produced by the organ anyway).



The second stage clearly shows my mod to improve the high end. The drop after 2kHz is gone, and in fact the goes up slightly instead. The response is now flat within 4dB from 50Hz to 20kHz, so anything fed into the line input will be reproduced with reasonable fidelity.