Presenting "The Leslieator" - a work in progress

Where we discuss new analog design ideas for Pro Audio and modern spins on vintage ones.
tubegeek
Posts: 25
Joined: Mon Mar 18, 2013 12:36 pm
Location: Brooklyn NY

Presenting "The Leslieator" - a work in progress

Post by tubegeek » Mon Jun 17, 2013 11:40 pm

On Sundays I mix sound for a church and, while my setup is quite functional, I'm always trying to do a little more with a little less.

Case in point: I currently have a stereo mic setup on a very nice Leslie 122 cabinet (driven, just as it is ordained in the scriptures, by a very nice Hammond B3.) The resources I use are: three mics, three mixer channels and one mix subgroup. Too much for one instrument, as lovely as it is. I have used small electret condenser mics in my day job and there are many circuits that show how to use these mic elements for various purposes, so I started developing this pre-mixer to economize on mixer resources.

My goal with this "Leslieator" is to pre-mix a "Left" pair of electrets, crossed over at 800 Hz, and also to do the same with a "Right" pair. Each of the pairs will have one mic for the top half of the Leslie cabinet and one for the bottom half. Another goal, equally as important, is to learn how to design and build such a thing. The blend and EQ are all just preset with equal gain top and bottom. The 800 Hz second-order crossover point matches the crossover frequency used inside the Leslie to split the signal before feeding the split into the upper and lower rotating drivers. (The drivers don't actually rotate, but you know what I mean.) (Possible future features: variable top/bottom balance, also experimenting with other crossover filter choices.)

So, a simple 4-into-2 mixer, with preset EQ and level. Two channels into my main mix, and no need for further separate signal paths for the top and bottom of the Leslie. Full stereo swirl on both top and bottom. (Possible future feature: mono blend for the two bottom mics.)

The story so far: it's built and almost working, using 4558's on Radio Shack proto board in a small project box.

By "almost," I mean that all 4 mics are showing up where they should in the mix, but there is also an oscillation present in both channels, even when the project box is closed up, shielding the circuit.

In response to this problem I added the 230 pF and 560 pF caps seen in the feedback loops of the two op-amp stages configured for gain. This did not kill the oscillation. The 1000pF caps shown in the mic detail are soldered right to the pins on the electret capsule. The electret capsule is connected to the box by two-conductor, shielded cable, and the capsules are tucked into the covers from 1/4" connectors which make for a shielded "body."

Power supply bypass is made from V+ to V- directly under each 4558, 10 nF ceramics soldered to the bottom of the board. Layout is not too ugly as far as I can tell. I will post a build shot soon.

Do the mic cables need to be shielded for the last few inches (inside the box)? At the moment the cable shields end at a small sub-board that carries the mic wiring, and the shields all connect to audio ground there. Then the individual mic signals travel a couple of inches to their op-amp +ve inputs on a single conductor apiece.

Any suggestions on how to stabilize this circuit? I'm SOOO close!

(By the way, ignore the cap and resistor values as shown around the filter sections, they are set up for 1KHz as shown, but I've recalculated them for 800Hz. I recently learned that this is the correct crossover frequency used in the Leslie.)

EDIT: Schematic has correct filter components for 800 Hz and I have removed the R's shunting the virtual earth nodes to ground per JR's correction.

Thanks,

-j
Attachments
Leslieator p1 edit.jpg
Leslieator schematic (rev 2)
Leslieator p1 edit.jpg (360.92 KiB) Viewed 13892 times
Last edited by tubegeek on Tue Jun 18, 2013 4:02 pm, edited 1 time in total.

User avatar
JR.
Posts: 3023
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: Presenting "The Leslieator" - a work in progress

Post by JR. » Tue Jun 18, 2013 10:20 am

That schematic looks OK from a quick glance.

I would lose the 1.3k resistors to ground on U5a and b,

Look for a wiring mistake or other error.

If you don't have a scope perhaps make a sniffer so you can listen the audio quality at different parts of the circuit to see where it gets messed up.

Or you can disconnect circuit sections to help isolate the problem.

JR
Don't only half-ass tune your drums. Visit https://circularscience.com to hear what properly "cleared" drums sound like.

tubegeek
Posts: 25
Joined: Mon Mar 18, 2013 12:36 pm
Location: Brooklyn NY

Re: Presenting "The Leslieator" - a work in progress

Post by tubegeek » Tue Jun 18, 2013 2:29 pm

JR. wrote:That schematic looks OK from a quick glance.
OK, thanks for the check-over.
JR. wrote:I would lose the 1.3k resistors to ground on U5a and b
Oops! Thanks! I was making a standard feedback configuration, not a virtual earth mixer stage configuration. Duh. My First Mixer™.
JR. wrote:Look for a wiring mistake or other error.
Now, how exactly would that even be possible? **I** wired it **myself!** ;)
JR. wrote:If you don't have a scope perhaps make a sniffer so you can listen the audio quality at different parts of the circuit to see where it gets messed up. Or you can disconnect circuit sections to help isolate the problem.
JR
Yeah, I have a scope but - get this - I've mislaid my frickin' probes somewhere! But there's a package on its way from Singapore with new ones that USPS tried to deliver yesterday, so, I'll be able to scope it out thoroughly, hopefully tonight or the next couple days. My observations so far have come from a quick once-over on the scope at work, but it was a pretty hasty check out so far.

Thanks for the help. Is there a recommended "correct" procedure to go through for sizing compensation caps like these, or is it usually basically a cut & try routine?

-j

User avatar
JR.
Posts: 3023
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: Presenting "The Leslieator" - a work in progress

Post by JR. » Tue Jun 18, 2013 3:02 pm

Those opamps are unity gain stable so in theory do not require any feedback caps. The opamps are internally compensated to be stable. In practice if there is some stray capacitance at the opamps - input that will cause some lag (phase shift) to the feedback signal that could make it unstable. Adding some lead capacitance across the feedback resistor will wash out the lag from stray capacitance.

I suspect you already have enough feedback C, if you will have scope probes soon, you can be more objective to see where the problem is.

JR
Don't only half-ass tune your drums. Visit https://circularscience.com to hear what properly "cleared" drums sound like.

tubegeek
Posts: 25
Joined: Mon Mar 18, 2013 12:36 pm
Location: Brooklyn NY

Re: Presenting "The Leslieator" - a work in progress

Post by tubegeek » Tue Jun 18, 2013 3:58 pm

Thanks!

I have just nipped the mistaken R's, and shot a couple of pics of the build-up. Also the schematic is now corrected for 800 Hz and no R's shunting the virtual earth nodes.

I think my fb C's are probably at least a decade too large but ....

Thanks!
Attachments
2013-06-18 16.13.44.jpg
inside the box
2013-06-18 16.13.44.jpg (88.77 KiB) Viewed 13892 times
2013-06-18 16.13.35.jpg
top of board
2013-06-18 16.13.35.jpg (125.68 KiB) Viewed 13892 times
2013-06-18 16.13.03.jpg
bottom of board w/power bypasses and FB caps
2013-06-18 16.13.03.jpg (125.79 KiB) Viewed 13892 times
2013-06-18 16.09.22.jpg
outside of the box
2013-06-18 16.09.22.jpg (113.59 KiB) Viewed 13892 times

tubegeek
Posts: 25
Joined: Mon Mar 18, 2013 12:36 pm
Location: Brooklyn NY

Re: Presenting "The Leslieator" - a work in progress

Post by tubegeek » Mon Jun 24, 2013 3:58 pm

An update:

My oscilloscope probes came in and I had a bit of time today to test the "Leslieator" out. It's working just fine and the oscillations went away as soon as I attached a clip lead between the body of the unit and my scope's ground connection. It just wasn't happy floating without a drain path, is all. All of the mic pres, filters, and mixers are stable without feedback compensation caps and I could see the response rolling off everything above about 8 - 10K - which seemed a bit excessive in terms of bandwidth restriction. So I removed all those caps. I may put back pF caps in their place but the values I had were WAY overkill. Not only that, they didn't do anything to fix the original problem!

So much for the shotgun approach.

I'll try it out next Sunday and see how it actually sounds - can't wait!

Thanks for the help everyone.

-j

tubegeek
Posts: 25
Joined: Mon Mar 18, 2013 12:36 pm
Location: Brooklyn NY

Re: Presenting "The Leslieator" - a work in progress

Post by tubegeek » Wed Jul 03, 2013 5:21 pm

Update:

Harsh, distorted sound. Accidentally I forgot to pan the two channels apart so I don't know if I heard a blend of one clean channel with one distorted, or what exactly. Looking at the scope, one channel was passing a strange waveform compared to the other. It's not exactly easy to get a pure sine wave acoustically, to use as a source into the mics, so it's a little hard to track everything down.

I am wondering if I am having a problem with the mics because I am feeding them such a high voltage: 15V is much more than the capsules are usually spec'ed for.

I plan to throw a 9V regulator on the V+ for the mic section and see if that makes a difference. Of course, I may have already ruined the capsules.

More updates as I get more info.

tubegeek
Posts: 25
Joined: Mon Mar 18, 2013 12:36 pm
Location: Brooklyn NY

Re: Presenting "The Leslieator" - a work in progress

Post by tubegeek » Sat Jul 06, 2013 10:39 pm

I installed an LM317 to regulate the mics' power node. It's adjusted to 7.8 V (the data sheets I've seen on similar electret capsules call for a max of 10V, and I didn't have the exact resistors handy to get 9V on the LM317) and the system is behaving perfectly now - at least, according to the 'scope. Anyway I don't think I damaged the capsules.

I'll try it on the Leslie speaker again tomorrow - I think I may have it this time, we'll see.

tubegeek
Posts: 25
Joined: Mon Mar 18, 2013 12:36 pm
Location: Brooklyn NY

Re: Presenting "The Leslieator" - a work in progress

Post by tubegeek » Sun Jul 07, 2013 11:57 pm

The circuitry is working perfectly, thanks for the help with it, Pros! Now it's just a matter of tweaking.

With the electret omnis, located just outside the Leslie cabinet's louvers, I get a lot of leakage from my other sources. (I had been using dynamic mics tucked INSIDE the Leslie cabinet previously.) I'll try placing the mics inside again - my only worry with that is the possibility of too much level being sent to the mic/line adjustable inputs on the mixer, it's a pretty hot signal as it is and I have the trims all the way down on the mixer. I may have to lower the gain on the mic preamp sections a little bit.

These mics have what seems to be better frequency range than the dynamics (as expected.) Overall it sounds pretty good, as long as I am careful not to get too much "wind noise" as the rotors go by. I added some foam windscreen material wrapped around the capsules, but I may need another layer on the bottom mics. Overall I'm quite pleased, it's a cool little box and I'm learning a lot with it.

User avatar
JR.
Posts: 3023
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: Presenting "The Leslieator" - a work in progress

Post by JR. » Mon Jul 08, 2013 10:48 am

Running the mics from full rated voltage will buy a couple dB more headroom. If you are getting too much level inside you can perhaps make some DIY ear plugs for them out of foam. It will probably scrape off some HF response but that may still sound OK.

Good luck.
Don't only half-ass tune your drums. Visit https://circularscience.com to hear what properly "cleared" drums sound like.

Post Reply