EMRFD Message Archive 9993
Message Date From Subject 9993 2014-04-30 18:54:27 bob_ledoux How Do You Attack Audio Amp Motorboating? I've built a number of audio amps, even from existing circuits, and seem to have problems with motorboating.
If I Google the issue I read individual exploits and the black magic that was involved in a particular problem. I'm looking for a general schema for locating the causes, isolating the source of the feedback, measurement methods, and problem solving.
Does a scope have utility here?
What about circuit board layout techniques?
Are there solution techniques, like poking a high value cap into the circuit until the problem ceases?
Are there articles that address a general set of guidelines for this issue?
9995 2014-04-30 19:44:46 Russell Shaw Re: How Do You Attack Audio Amp Motorboating? 9996 2014-05-01 06:15:31 bob_ledoux Re: How Do You Attack Audio Amp Motorboating? 9997 2014-05-01 06:35:17 Nick Kennedy Re: How Do You Attack Audio Amp Motorboating? I recall a warning that motorboating might occur with the Norcal Tin Ear receiver, IIRC, if you used a "heavy duty" 9V battery instead of an alkaline. I think this receiver uses a discrete version of the LM386 as is shown in EMRFD.I did some measurements and found that an alkaline 9V battery had an internal resistance of 1.5 ohms but a heavy duty battery's internal resistance was 45 ohms.I believe putting a large value electrolytic across the lower quality battery (if used) might help with the problem.73-Nick, WA5BDU 9998 2014-05-01 09:40:49 Sandeep Lohia Re: How Do You Attack Audio Amp Motorboating? I was using LM386 with a 100 ohm resistance at it's positive supply pin no 6,
wth 10 ohm it used to motorboat...
9999 2014-05-01 12:29:38 arfghans Re: How Do You Attack Audio Amp Motorboating? Your most likely culprit is current feedback via the load. High current that circulates thru the load, if allowed to return to an arbitrary point near the input of the amplifier, can result in a small I*R drop (voltage) that is then coupled into the amplifier input. This is an unwanted feedback loop. What you want to do instead is return the load to the power supply common as directly as possible. Then you can follow the circuit for high current and it is very clear: From PS+ to the amplifier, out of the amplifier, thru the load, and back to PS-. Low-level signals cannot intermix with this high-current loop, ever! The input signal common for the amplifier is also PS-, but via a separate conductor path. Power supply return is always a single-point ground in a well-designed audio power amplifier system.Gary, WB9JPS 10000 2014-05-01 15:58:16 bob_ledoux Re: How Do You Attack Audio Amp Motorboating? The circuit is my SMT layout of the Norcal 2030 receiver audio stage. It consists of a quad TLV2464 op amp. The first amp is a three pole high pass filter. The second stage is a 43 db audio amp. The third and fourth amps are parallel drivers for earphones. Before those last two amps is a MAX7427 SCAF filter. This is the circuit found on page 3 of the Norcal published schematic document.
I've taken the proven circuit from my 2030 and redrawn the diagram slightly larger with my components moved a bit. An LP2985-2.9 volt regulator drives the audio portion. One side of the board has the etched circuit and the other side is a ground plane. Vias connect front to back when a ground is indicated. The ground plane is continuous. There are no ground traces on the trace side of the board.
I assumed a ground layer in a circuit board qualifies as a single point ground. So this is in error. But I don't see a single point ground in the original 2030 board.
The circuit has a 100uf bypass cap at the op amp power supply and four additional 22uf bypass caps at various points in the circuit.
The motorboating exists at high gain levels. In my current test configuration I have 6 inch wires to the volume control that might contribute to the problem.
I didn't expect a proven circuit to be this sensitive to layout changes.
10001 2014-05-01 16:27:31 David Re: How Do You Attack Audio Amp Motorboating? Ground planes do not count as single point grounds if the gain is high enough or
the common mode or power supply rejection is low enough. In some cases it is
necessary to cut gaps in a ground plane to route high currents away from
On 01 May 2014 15:58:16 -0700, you wrote:
>I assumed a ground layer in a circuit board qualifies as a single point ground. So this is in error. But I don't see a single point ground in the original 2030 board.
10002 2014-05-01 19:04:53 vasilyivanenko Re: How Do You Attack Audio Amp Motorboating?
Also.... decouple plus bypass to filter the DC lines in your various AF stages, especially the AF power amp.
Expect AF feedback motorboating via your DC power lines as you increase the AF gain pot in many LM386 circuits. If this happens, try better bypass and decoupling on the LM386 power line.