iraudamp1 problem

[webmans]

I am testing my own layout for the iraudamp1, having followed their schematic exactly. I am using all the same components also. I wanted to do this first to get something working so I can understand the thing and then go on to modifying the design. I could not get self-oscillation at 300khz with the original component values. It maxed out at about 200. I assume this means my layout is causing more phase shift than theirs. I lowered the integrator caps to 510pf instead of 1000. Now it is oscillating at 340kHz and the switching waveform looks fine; good square wave with no significant ringing. But if I feed a sine wave audio signal, the output starts to become a triangle wave at about 7V, irrespective of the load. At first I thought it must be choke saturation, but I now have a large RM core choke in there and it is definitely not saturating and the waveform starts to distort at the same level. For testing I am using +/- 35V regulated bench supply. I noticed there were several people with experience of this amp on the forum. Any ideas what I should check?

[zkaiser]

are u sure that your choke is properly wound and that has needed inductance ~18uH, where did you put 470u caps, and post some hires. picture of device . . .

BTW you can use 5k pot. for frq. adj. to make room for wider phase comp. which will in the end make possible oscill. at higer freq.

[webmans]

Thanks for the input. Attached is a picture of the amp as it is now...only one channel parts are in and the heatsink is just a temporary one as is the mounting of the choke. The layout is for a gapped toroid but I don't have one yet. I've also attached my copy of the schematic...as I said as far as I know it follows the IR. Also including the top silk screen layer to show where the parts are placed. One thing I should say, we are doing this over in India as a education project (I'm teaching some guys) and critical parts are not easily available (like the high speed mosfets...), so I take a lot of measures not to blow anything during the development stage. We are running off a bench supply with selectable current limiting. The large (10,000 mfd) electrolytics outside the board were necessary because of the long line from the supply to the board under test.

I naturally suspected the choke first, but can't find anything wrong. We have a meter and it is 15microhenry (IR said 10 but I realized that was too low). We made one like this for an SMPS with 50V out loaded at 3A and it worked fine. As you see there is ample room to add a few turns which I will do. But the problem of the audio sine wave becomming a triangle wave seems to be independent of the choke. I had a 10 microhenry commercial choke for the first test and the only difference was that there was too much carrier coming through the filter. Moreover that choke probabably was near saturation. But the waveform distortion started at the same voltage (about 7V). And it is the same irrespective of the load...open, 8 ohms or 4 ohms, so it is not current related.

Tomorrow I will try your suggestion about the integrator, and see if that somehow helps.

[webmans]

here is the schematic...as I said straight copy of the IR circuit

[webmans]

here is the component layout, couldn't fit the pdf so I changed it to jpg

[crriss]

First of all, where u putted the 470uF cap. from what i see in that picture, it seems to be a little bit far from the transistors. u should put them as close as possible to the transistors, and also try to put some non pol. capacitors, 100-470nF, solder them dirrectely on to transistors legs. I had a simillar problem with a class T amp, and i solve the problem by just putting the caps. very close to the IC.

[webmans]

thanks. It appears that the problem was in fact the lower capacitors on the integrator. Taking zkaiser's suggestion I replaced the 1K trimmer with 5K and put back the 1000pf caps and now the output looks OK. I will increase the choke turns a little now as there is now more carrier coming through, and will also put bypass caps directly on the pins of the MOSFETs as suggested by Crriss. Since the position of the 470 UF caps didn't seem to be the basic cause of my problem, I'll only change that in my next layout. Thanks for the help.

[luka]

What can be used instead of LT1220 ?

[webmans]

I've got a couple of Texas Instruments THS4051 that at first glance at least, seem to have similar specs. I haven't tested them yet and it will be a few weeks before I can work on this project again. The TI op amps are lower cost than the LT1220 and easier to find.

[luka]

Will LM(118,218 or 318) do the job?