A&T K6 Power Amplifier - homebuilt

BlueCrab

Journeyman
Joined
Nov 10, 2019
Messages
229
#1
So almost 2 years ago I was scouting around for something to build - something really different other than a restoration. I began looking into a bunch of old electronics magazines (found here: World Radio History) and eventually came upon this: K6 Amplifier. What intrigued me about this was the SMPS power supply. To digress for a moment, most of my career was spent in avionics where SMPSs are heavily used due to the weight savings. So I had some experience. This looked like an opportunity to build a big amp, but not give me a hernia every time I wanted to move it. Plus the amp itself was interesting - power MOSFETs. Now this is an old design - early to mid 90's. I contacted Reinhart Metz - the designer and owner - he still had some parts for the SMPS, not the pulse transformers, but had a suggestion for a commercially available suitable sub. So I ordered a shitload of parts from Mouser, Digikey, Newark, EBay, and others.

I started on the SMPS. I turned it on - no smoke - voltmeter readings looked ok. And then I looked at the plus & minus output with a scope. There was this 2 V p-p ripple at roughly 8 Hz. Long story shorter, I futzed with this for a year. I tried this, I tried that, maybe this will fix it. I tried breadboarding it. I bought every pulse transformer I could (they are not expensive). No joy. Damn near pitched it out the window. Finally I gave up. Put it in a box and set it on my shelf.

So I switched to building the amp and built a conventional power supply with a toroidal transformer. Got both amplifier boards together and turned them on. They made nice 1.5MHz oscillators (roughly 1.5MHz, each had its own special freq. We use to say if you want an oscillator, build an amplifier.). Ugh. Finally tamed these beasts - turned down the loop gain and reduce the overall bandwidth (originally it was about 750KHz - not much musical content there). Mounted the boards, heat sink, conventional supply in a chassis - weighs about 40lb - gives me a hernia every time I move it. But it works - nicely. I've been using it (listening to it now) regularly for the last six months. (There's more to this story, but I'm cutting it short - let's just say "hum".)

Some time ago on this forum someone posted a link about capacitors. I learned that at some frequencies some capacitors look more like inductors than capacitors and it got me thinking that my SMPS filter capacitors are trying to filter 77kHz. One of the improvements I made to the SMPS was to use 200vdc electrolytic capacitors instead of the spec'd 100vdc caps. Investigating this further I discovered that SMPS in particular are sensitive to the filter caps and you should use low ESR caps. New order to Mouser and new caps installed and the SMPS works. At least to the loads I'm able to test - 150 ohm load @ 80vdc = 43 watts x 2 (plus and minus 80vdc) = 86 watts. It's supposed to be a 1kw power supply.

IMG_1786.JPEG
IMG_1784.JPEG IMG_1787.JPEG
 

Gepetto

Administrator
Staff member
Joined
May 15, 2011
Messages
13,473
Location
Sterling, MA
Tagline
Old 'Arn Enthusiast
#2
So almost 2 years ago I was scouting around for something to build - something really different other than a restoration. I began looking into a bunch of old electronics magazines (found here: World Radio History) and eventually came upon this: K6 Amplifier. What intrigued me about this was the SMPS power supply. To digress for a moment, most of my career was spent in avionics where SMPSs are heavily used due to the weight savings. So I had some experience. This looked like an opportunity to build a big amp, but not give me a hernia every time I wanted to move it. Plus the amp itself was interesting - power MOSFETs. Now this is an old design - early to mid 90's. I contacted Reinhart Metz - the designer and owner - he still had some parts for the SMPS, not the pulse transformers, but had a suggestion for a commercially available suitable sub. So I ordered a shitload of parts from Mouser, Digikey, Newark, EBay, and others.

I started on the SMPS. I turned it on - no smoke - voltmeter readings looked ok. And then I looked at the plus & minus output with a scope. There was this 2 V p-p ripple at roughly 8 Hz. Long story shorter, I futzed with this for a year. I tried this, I tried that, maybe this will fix it. I tried breadboarding it. I bought every pulse transformer I could (they are not expensive). No joy. Damn near pitched it out the window. Finally I gave up. Put it in a box and set it on my shelf.

So I switched to building the amp and built a conventional power supply with a toroidal transformer. Got both amplifier boards together and turned them on. They made nice 1.5MHz oscillators (roughly 1.5MHz, each had its own special freq. We use to say if you want an oscillator, build an amplifier.). Ugh. Finally tamed these beasts - turned down the loop gain and reduce the overall bandwidth (originally it was about 750KHz - not much musical content there). Mounted the boards, heat sink, conventional supply in a chassis - weighs about 40lb - gives me a hernia every time I move it. But it works - nicely. I've been using it (listening to it now) regularly for the last six months. (There's more to this story, but I'm cutting it short - let's just say "hum".)

Some time ago on this forum someone posted a link about capacitors. I learned that at some frequencies some capacitors look more like inductors than capacitors and it got me thinking that my SMPS filter capacitors are trying to filter 77kHz. One of the improvements I made to the SMPS was to use 200vdc electrolytic capacitors instead of the spec'd 100vdc caps. Investigating this further I discovered that SMPS in particular are sensitive to the filter caps and you should use low ESR caps. New order to Mouser and new caps installed and the SMPS works. At least to the loads I'm able to test - 150 ohm load @ 80vdc = 43 watts x 2 (plus and minus 80vdc) = 86 watts. It's supposed to be a 1kw power supply.

View attachment 72443
View attachment 72444 View attachment 72446
For stability reasons, just about all SMPS power supplies require a minimum high frequency ESR (in other words the cap cannot be too good) which forms a zero in the network at high frequency to cancel one of the 2 poles in the LC output filter. If you use a cap that is too good, you will get oscillation. If you use a cap that cannot tolerate the high frequency ripple current, you will get a bulging cap which overheats and eventually fails.
 

BlueCrab

Journeyman
Joined
Nov 10, 2019
Messages
229
#3
There were a bunch of threads concerning this power supply back in the 05 - 06 years where folks were having problems with stability. No one ever mentioned (or I missed it - could be true too) the critical nature of the filter caps. I spent a lot of time investigating the PI filter network, thinking the issue was there. Learned a lot about PI and PID networks though. I should have contacted Reinhart, he probably would have recognized the problem.
 

Mohawk

Chief Journeyman
Joined
Oct 29, 2019
Messages
588
Location
Ontario Canada
#5
So almost 2 years ago I was scouting around for something to build - something really different other than a restoration. I began looking into a bunch of old electronics magazines (found here: World Radio History) and eventually came upon this: K6 Amplifier. What intrigued me about this was the SMPS power supply. To digress for a moment, most of my career was spent in avionics where SMPSs are heavily used due to the weight savings. So I had some experience. This looked like an opportunity to build a big amp, but not give me a hernia every time I wanted to move it. Plus the amp itself was interesting - power MOSFETs. Now this is an old design - early to mid 90's. I contacted Reinhart Metz - the designer and owner - he still had some parts for the SMPS, not the pulse transformers, but had a suggestion for a commercially available suitable sub. So I ordered a shitload of parts from Mouser, Digikey, Newark, EBay, and others.

I started on the SMPS. I turned it on - no smoke - voltmeter readings looked ok. And then I looked at the plus & minus output with a scope. There was this 2 V p-p ripple at roughly 8 Hz. Long story shorter, I futzed with this for a year. I tried this, I tried that, maybe this will fix it. I tried breadboarding it. I bought every pulse transformer I could (they are not expensive). No joy. Damn near pitched it out the window. Finally I gave up. Put it in a box and set it on my shelf.

So I switched to building the amp and built a conventional power supply with a toroidal transformer. Got both amplifier boards together and turned them on. They made nice 1.5MHz oscillators (roughly 1.5MHz, each had its own special freq. We use to say if you want an oscillator, build an amplifier.). Ugh. Finally tamed these beasts - turned down the loop gain and reduce the overall bandwidth (originally it was about 750KHz - not much musical content there). Mounted the boards, heat sink, conventional supply in a chassis - weighs about 40lb - gives me a hernia every time I move it. But it works - nicely. I've been using it (listening to it now) regularly for the last six months. (There's more to this story, but I'm cutting it short - let's just say "hum".)

Some time ago on this forum someone posted a link about capacitors. I learned that at some frequencies some capacitors look more like inductors than capacitors and it got me thinking that my SMPS filter capacitors are trying to filter 77kHz. One of the improvements I made to the SMPS was to use 200vdc electrolytic capacitors instead of the spec'd 100vdc caps. Investigating this further I discovered that SMPS in particular are sensitive to the filter caps and you should use low ESR caps. New order to Mouser and new caps installed and the SMPS works. At least to the loads I'm able to test - 150 ohm load @ 80vdc = 43 watts x 2 (plus and minus 80vdc) = 86 watts. It's supposed to be a 1kw power supply.

View attachment 72443
View attachment 72444 View attachment 72446
Love the Marantz 150 , I have one of those beauties myself .... Not to common that one !
 

BlueCrab

Journeyman
Joined
Nov 10, 2019
Messages
229
#6
Restored the Marantz about 6-8 years ago. Also have a Yamaha CT-7000 tuner. BTW, FM is the way I stream most of my music around the house. Or actually I stream music to a FM transmitter that I tune my FM receivers to.
 
Top