TGM6 amplifier (7 channels) - diyAudio
Go Back   Home > Forums > Amplifiers > Solid State

Solid State Talk all about solid state amplification.

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 20th February 2013, 12:48 AM   #1
Bigun is offline Bigun  Canada
diyAudio Member
 
Bigun's Avatar
 
Join Date: Jan 2009
Blog Entries: 2
Default TGM6 amplifier (7 channels)

A few weeks ago I pulled an old Pioneer amp out of the garbage. I pulled out all the boards except the amp and power supply, pulled out all the controls and junk off the front panel.

I was initially intrigued by the amplifier design, it has a lot in common with my first project, the TGM (TGM1) amplifier - Pioneer VSX1020 like TGM1 ?

After taking some time to find my around the schematics I was able to get it powered up. With most of the boards scrapped I had to install some jumpers, fix up the grounding scheme, enable the output speaker relays and wire in a temporary volume control on one of the channels.

The original amp was set up with a design target of 2mV across the output emitter resistors. Being a fan of D. Self I adopted the more acceptable target of closer to 25mV.

I didn't want to waste any more time until I'd heard it so I decided today to fire it up. And now that I've heard it, I think it has potential and so I'll be spending a bit more time on this design.

For my initial test I used a PMC speaker, a Sony blu-ray player and a YBA CD player for source material (Pink Floyd Animals and Christy Baron).

Initially I got some hum in the speaker and wondered if my earthing scheme was messed up but it was eliminated by powering off my nearby TGM5 amplifier; the TGM5 has a power supply that leaks flux like nobody's business.

The sound was cleaner from the YBA than from the Sony player. I used my Bryston BP60 amp as a reference amp, with the left channel from the source playing through the Pioneer amplifier and the right channel playing through the Bryston amplifier. The difference between them was not very significant - in fact I was a bit surprised by how good the Pioneer sounded. Ultimately the Bryston sounded cleaner, but the Pioneer amp lost little in terms of bass, treble and overall dynamics. It did have big power on/off thumps though due to how I've set up the power to speaker output relays (speaker protection is enabled and I tested it beforehand using a d.c. source).

So I will do some more cleaning up and consider some further modifications to see if this can become my next generation, "TGM6", amplifier.

The attached image shows the TGM6 making it's first sound. I've used one of the original pcb's from the amp as a platform to provide RCA inputs and a convenient connection to the amplifier inputs. There's no power to this pcb and so the parts on it are non-functional. I removed some coupling capacitors to provide convenient through-holes to install the blue coloured Cat5 cables you can see. These cables run to the front panel and provide connection for a passive volume control pot (50k Ohm). Sitting inside and on the bottom of the chasis is the power supply pcb along with the output speaker protection relays. You can't see the amplifier pcb here, just the tip of the back of the amplifier but you can see the fins from the heatsink.
Attached Images
File Type: jpg first sound.JPG (440.5 KB, 375 views)
__________________
"The test of the machine is the satisfaction it gives you. There isn't any other test. If the machine produces tranquility it's right. If it disturbs you it's wrong until either the machine or your mind is changed." Robert M Pirsig.

Last edited by Bigun; 20th February 2013 at 12:50 AM.
  Reply With Quote
Old 20th February 2013, 01:05 PM   #2
Bigun is offline Bigun  Canada
diyAudio Member
 
Bigun's Avatar
 
Join Date: Jan 2009
Blog Entries: 2
Taking a look at the amplifier: It's a LIN topology, LTP input. It simulates at around 0.01% THD when driving +/-30V into an 8 Ohm resistive load - although I don't have the spice models for the Darlington output devices and used a discrete model instead. The FFT profile is typical and unremarkable.

The LTP has an independent, regulated power supply. This improves PSRR and reduces channel-channel cross-talk. A resistor connects the output to the top of the LTP, modulating the current in step with the signal applied to the bases of the LTP transistors. In theory, like a Gilbert Cell mixer, this results in some creation of harmonic distortion, primarily 2nd order.
Another interesting feature is lack of Cdom phase lag compensation at the VAS. Instead there is resistive load added to the VAS collector and emitter degeneration, plus a resistor + cap from the VAS base to the -ve rail. There is some additional compensation, in the form of phase lead at the feedback node.

If I'm not mistaken, the input is also bootstrapped so input impedance is dominated by the 100K resistor to ground.
Attached Images
File Type: jpg TGM6.jpg (113.3 KB, 327 views)
File Type: jpg FFT.jpg (43.7 KB, 311 views)
Attached Files
File Type: asc Pioneer.asc (20.5 KB, 9 views)
__________________
"The test of the machine is the satisfaction it gives you. There isn't any other test. If the machine produces tranquility it's right. If it disturbs you it's wrong until either the machine or your mind is changed." Robert M Pirsig.
  Reply With Quote
Old 16th March 2013, 11:24 PM   #3
Bigun is offline Bigun  Canada
diyAudio Member
 
Bigun's Avatar
 
Join Date: Jan 2009
Blog Entries: 2
I have built-in some additional circuitry to ensure that on power-up the rails to the amplifier come up first, then the speaker relays close after about 8s of delay. On power-down the speaker relays open quickly and the rails to the amplifier take nearly a minute to collapse.

Next I am exploring some changes to the amplifier. The basic design has relatively low nfb so there's a fair bit of differential signal at the LTP. This leads to the problem of the LTP transconductance drooping on large signal swings. Adding degeneration to the LTP flattens the transconductance but the gain drops rapidly so that in the end there's no benefit. A singleton input would have thermal drift issues that I don't want to deal with in this design. What I need is higher OLG gain.

The common solution is a current mirror. I don't want to use a current mirror, it would be difficult to install on the pcb and for various reasons I'm not a fan of them. So how else to reduce the differential signal at the LTP ?

I am looking at two options

1) bootstrap the collector load of the LTP to increase gain. See 1st attachment. There's a useful signal at the emitter of the VAS due to degeneration and as this is a low impedance node it can be used to bootstrap the LTP collector.

2) in addition to 1) I will try a CFP for the VAS. With Cdom compensation in the regular Lin topology the CFP doesn't offer so much benefit since Cdom provides local feedback. But this amp doesn't use regular phase lag compensation so a CFP offers a benefit. It also provides a better buffer between the high impedance bootstrapped LTP and the output devices.
Attached Images
File Type: jpg TGM6v1.jpg (66.5 KB, 255 views)
File Type: jpg TGM6v2.jpg (69.7 KB, 259 views)
__________________
"The test of the machine is the satisfaction it gives you. There isn't any other test. If the machine produces tranquility it's right. If it disturbs you it's wrong until either the machine or your mind is changed." Robert M Pirsig.

Last edited by Bigun; 16th March 2013 at 11:30 PM.
  Reply With Quote
Old 17th March 2013, 09:37 PM   #4
Bigun is offline Bigun  Canada
diyAudio Member
 
Bigun's Avatar
 
Join Date: Jan 2009
Blog Entries: 2
Today I built up TGM6.0 (as per original Pioneer topology), TGM6.1 (bootstrapped LTP) and TGM6.2 (with CFP Vas) - since I have 7 channels I can build them all and compare.

The CFP VAS had to be a pin-compatible replacement for the existing VAS transistor. So some point-to-point wiring was needed to construct a CFP VAS device. See photos.
Attached Images
File Type: jpg CFP.JPG (34.8 KB, 135 views)
File Type: jpg CFP from back.JPG (29.7 KB, 71 views)
__________________
"The test of the machine is the satisfaction it gives you. There isn't any other test. If the machine produces tranquility it's right. If it disturbs you it's wrong until either the machine or your mind is changed." Robert M Pirsig.
  Reply With Quote
Old 17th March 2013, 09:57 PM   #5
wahab is offline wahab  Algeria
diyAudio Member
 
Join Date: Nov 2009
Location: algeria/france
Quote:
Originally Posted by Bigun View Post
The common solution is a current mirror. I don't want to use a current mirror, it would be difficult to install on the pcb and for various reasons I'm not a fan of them. So how else to reduce the differential signal at the LTP ?
Implementing a current mirror will somewhat improve the things
but not that dramaticaly and it will be the same if the VAS is
enhanced , only by implementing thoses two mods together
there will be a valuable improvement.
  Reply With Quote
Old 17th March 2013, 11:44 PM   #6
Bigun is offline Bigun  Canada
diyAudio Member
 
Bigun's Avatar
 
Join Date: Jan 2009
Blog Entries: 2
Well, it would require more surgery on the pcb to implement a mirror. However, all is not lost since with the bootstrap the impedance of the LTP is increased a lot - not as much as a mirror but still higher. With a full mirror it may require more compensation which I'd like to avoid. Then there's the distortion profile - some say 2nd harmonic is benign. A mirror balances the LTP and reduces mostly 2nd harmonic.

I'm sure I could support either solution, but I like the simple bootstrap - which doesn't use a capacitor so it's direct coupled. The input of the LTP is also bootstrapped off the feedback node. And the VAS is bootstrapped. No shortage of boots then.

Measurements with square waves indicated more rounding than I would like to see at 10kHz so I removed the 10pF Cdom from the VAS of 6.1 and 6.2. I didn't find any r.f. instability issues with the CFP.

Listening impressions: the bootstrap isn't working well - it measures terribly with lots of distortion and it sounds like it. Perhaps I have a wiring error as the sims were not this bad. With the LTP bootstrap + CFP Vas it appears to be working well. I'm hearing a cleaner sound with more punch in the bass. Treble is cleaner, cymbals that are splashy on v6.0 (pioneer) are clear on v6.2 (CFP VAS). Cello strings retain more realism with v6.2 (Hadyn). At high volumes with lots of bass the CFP version is clearly tighter. I am very sensitive to treble and apart from sibilance on both amps on one CD (Rod Stewart) I detect no issues with the CFP - good (with TGM2 amplifier with CFP in the LTP I has some concerns over the treble).

Clipping looks about the same between 6.0 and 6.2, both clip around +/-42V. It's not perfectly symmetrical but it's not far off. When I push it way hard beyond clipping I can see some sticking to the +ve rail - the waveforms are the same for both versions of the amp.
__________________
"The test of the machine is the satisfaction it gives you. There isn't any other test. If the machine produces tranquility it's right. If it disturbs you it's wrong until either the machine or your mind is changed." Robert M Pirsig.

Last edited by Bigun; 17th March 2013 at 11:50 PM.
  Reply With Quote
Old 18th March 2013, 12:01 AM   #7
wahab is offline wahab  Algeria
diyAudio Member
 
Join Date: Nov 2009
Location: algeria/france
Well , if that must stay simple then at least reduce the VAS emitter 47R
resistor to about 10-15R , it should reduce THD significantly.
  Reply With Quote
Old 18th March 2013, 01:10 AM   #8
Bigun is offline Bigun  Canada
diyAudio Member
 
Bigun's Avatar
 
Join Date: Jan 2009
Blog Entries: 2
Hi Wahab, yes it does seem high, a value as you suggest is more common, but the 47R serves some useful purpose. It provides the emitter load necessary to produce a bootstrap voltage for the LTP collector - increasing the effective load impedance and LTP gain considerably. It is also part of the compensation, in reducing the VAS transconductance it helps with closed loop stability. And with the CFP it has been my experience that you need more than 10 - 15R for the CFP to function well - it needs the collector of the slave device to be adequately isolated from the -ve rail so that it can feedback to the VAS emitter master device. All in all I don't think I can change the value.
__________________
"The test of the machine is the satisfaction it gives you. There isn't any other test. If the machine produces tranquility it's right. If it disturbs you it's wrong until either the machine or your mind is changed." Robert M Pirsig.
  Reply With Quote
Old 18th March 2013, 11:04 PM   #9
Bigun is offline Bigun  Canada
diyAudio Member
 
Bigun's Avatar
 
Join Date: Jan 2009
Blog Entries: 2
Measured small signal frequency response. Sine wave input, +/- 1V into 8 Ohm dummy load measuring amplitude of output using a scope and plotted in Excel. Horizontal axis is Log frequency (kHz), vertical axis is a linear scale, gain normalized to 1 at 2kHz.

I simulated it in spice and it doesn't match as well as I had expected. The measured results show worse low frequency extension but better high frequency extension.

As measured, it shows -3dB points at 4.5 Hz and 500kHz.
Attached Images
File Type: jpg TGM6v2FR.jpg (26.5 KB, 56 views)
__________________
"The test of the machine is the satisfaction it gives you. There isn't any other test. If the machine produces tranquility it's right. If it disturbs you it's wrong until either the machine or your mind is changed." Robert M Pirsig.

Last edited by Bigun; 18th March 2013 at 11:06 PM.
  Reply With Quote
Old 20th March 2013, 01:19 AM   #10
Bigun is offline Bigun  Canada
diyAudio Member
 
Bigun's Avatar
 
Join Date: Jan 2009
Blog Entries: 2
With setting a proper bias for the outputs the heatsink is going to get warmer than the original designers intended. So I've added a couple of low noise fans to the heatsink.

Square wave testing looking good.

And some distortion FFT plots from Spice showing a very healthy profile.
Attached Images
File Type: jpg add fans.JPG (109.4 KB, 78 views)
File Type: jpg squarewave.JPG (26.8 KB, 81 views)
File Type: jpg FFT2kHz.jpg (319.8 KB, 65 views)
__________________
"The test of the machine is the satisfaction it gives you. There isn't any other test. If the machine produces tranquility it's right. If it disturbs you it's wrong until either the machine or your mind is changed." Robert M Pirsig.
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Can I run tweeters off front channels and woofers off the rear channels of a headunit codyd70 Car Audio 0 3rd October 2010 06:11 AM
How can I wire my amplifier so I can bridge 2 channels together? soundengine355 Solid State 3 1st April 2009 09:13 PM
Combine 4 amplifier channels into one? thylantyr Solid State 5 4th March 2008 07:43 PM
Combining 2 channels to one klitgt Parts 16 22nd July 2005 05:24 PM
how to connect a sub on a 2 channels amplifier Mikka Multi-Way 1 19th June 2003 10:27 AM


New To Site? Need Help?

All times are GMT. The time now is 12:53 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2