TGM8 - an amplifier based on Rod Elliot P3a - 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 9th November 2013, 09:14 PM   #1
Bigun is offline Bigun  Canada
diyAudio Member
 
Bigun's Avatar
 
Join Date: Jan 2009
Blog Entries: 2
Default TGM8 - an amplifier based on Rod Elliot P3a

Rod's P3a has been popular and well regarded. Others have taken the basic amplifier and made their own improvements to it, some parts choices, some rail filters, even current mirrors. This is my take on it.

I want to retain the overall sound, the essence of the P3A but perhaps with some enhancements.

First thing to look at is the input stage. I'll look at the VAS and output stage later.

This is probably the most major change to the design, a change in topology from LTP with CCS feed, to a Singleton input. The Singleton has a different I-V curve from the LTP and with lowish feedback designs like this it can offer less compression with large signals without giving up anything in terms of OLG. The Singleton can be easier to stabilize, so Cdom can be smaller and slew rate is higher. The original P3a has no emitter degeneration on the LTP which means it is susceptible to IM distortion - we avoid that with the Singleton input.

However, the Singleton does not have good temperature stability and if you want to remove the input cap like I do, then you need to d.c. bias the feedback network to allow for the Vbe drop between base-emitter of the input device. I've used a zener diode to create a negative reference voltage independent of the supply rail voltage. From this I've created a voltage divider from the zener reference to ground using a resistor and a transistor wired as a Vbe multiplier. By adjusting the Vbe multiplier I can change the d.c. bias fed into the feedback network and thereby set the d.c. offset at the amplifier output. The Vbe multiplier that is part of this biassing provides temperature compensation for the input stage.

PSRR is good, distortion is at least as good as the P3a. I think it will preserve the flavour of the P3a and perhaps enhance the performance.

Update: July 2014: the amplifier, having been built, tested and listened to meets all my expectations. I'm very pleased with it and don't find anything that I wish to be improved on. Although inspired by P3a it also takes ideas from other many other talented designers, building on the limited wisdom I have collected from this forum and other internet sites and has now become a fully thought-out design in its own right. I can at least say that one listener has compared it with the P3a and says the TGM8 does retain the flavour of Rod's design and yet surpasses it....

And so, this amplifier also marks the end of my journey, as I see it, in the design and construction of SS amplifiers that started when I entered this hobby with the original TGM amplifier, based off AKSA 55. I find nothing 'wanting' with the TGM8, it runs with low idle current, supports difficult loads at good power levels, sounds clean, smooth and with great bass and dynamics. Yet has nothing to fatigue the listener, it is never harsh, cold or clinical. Anybody wanting to build this amp for DIY purposes should find all they need in this thread; the pcb files are in posted somewhere after post 399. Enjoy.
Attached Images
File Type: jpg tgm8-prelim.jpg (193.8 KB, 3722 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; 1st July 2014 at 10:41 PM.
  Reply With Quote
Old 9th November 2013, 10:02 PM   #2
diyAudio Member
 
Join Date: Feb 2010
Location: Rio de Janeiro
Very interesting.
  Reply With Quote
Old 10th November 2013, 01:02 AM   #3
AKSA is offline AKSA  Australia
diyAudio Member
 
Join Date: Sep 2001
Location: Melbourne, Australia
Fascinating!

Hugh
__________________
Aspen Amplifiers P/L (Australia)
www.aksaonline.com
  Reply With Quote
Old 10th November 2013, 01:26 AM   #4
MiiB is online now MiiB  Denmark
diyAudio Member
 
Join Date: May 2005
Location: Denmark
I like..
  Reply With Quote
Old 10th November 2013, 04:54 AM   #5
diyAudio Member
 
Join Date: Jul 2010
Location: Cacak
Can't wait to see some kind of pcb design for this amp.
  Reply With Quote
Old 10th November 2013, 02:28 PM   #6
Bigun is offline Bigun  Canada
diyAudio Member
 
Bigun's Avatar
 
Join Date: Jan 2009
Blog Entries: 2
I forgot to mention another key thing about the input stage. By controlling the d.c. level in the feedback network in this way, the feedback capacitor is quickly brought up to the correct bias voltage and doesn't create the turn-on issues normally associated with a Singleton input. There are other contributions to the turn-on behaviour of the P3a and this is something I still have to look at.

PCB ? --- I'm afraid my recent designs have been unpopular around here because I make use of a mixture of through-hole and surface mount. Some people really are still afraid, in this day, of surface mount. My plan so far on my TGM7 has worked fairly well. I use through hole for parts with 3 or more connections such as transistors and trim-pots. I use through hole parts for all large capacitors too. Where I have been using surface mount is for resistors and even then I use 1206 sized parts. These are parts are big enough to read their labels so you can read off the values of them. Have a look at the attached photo to see a through hole resistor lying on top of a pcb near a surface mount resistor. It's not hard to use at all, a pair of small pliers to hold it in place whilst you solder one end at a time. You end up being able to solder them in quickly and easily, with little solder. They allow for shorter signal traces and more compact design which can be very handy when your feedback amplifier has open loop unity gain in the MHz region. The only disadvantage is that you can't easily remove surface mount resistors unless you have two soldering irons, or be prepared to crack it and remove it in two pieces. I haven't decided which approach to designing the pcb to take with TGM8 yet.
Attached Images
File Type: jpg IMG_2964.JPG (101.7 KB, 3335 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; 10th November 2013 at 02:46 PM.
  Reply With Quote
Old 10th November 2013, 02:35 PM   #7
diyAudio Member
 
KatieandDad's Avatar
 
Join Date: Nov 2011
Location: UK
I use a croc clip with two flattened cocktail sticks to hold the surface mount components.
  Reply With Quote
Old 10th November 2013, 03:04 PM   #8
diyAudio Member
 
Join Date: Jul 2003
Location: wyoming
Smd 1206 size resistors are quite easy to solder and desolder if one uses a magnifying lamp. I put some solder on one pad, then put the resistor in place and hold it until the solder cools, 1-2 seconds. Easy to do the other side as the resistor is held in place by the first solder. Desoldering with braid is pretty much the reverse of the above. It's possible to save a desoldered resistor, but they are so cheap that I usually use new ones.
For me, using the magnifying lamp is essential and I now use it for all soldering.
  Reply With Quote
Old 10th November 2013, 06:52 PM   #9
Bigun is offline Bigun  Canada
diyAudio Member
 
Bigun's Avatar
 
Join Date: Jan 2009
Blog Entries: 2
Output Stage - I want to preserve the sound signature, so we keep the simple CFP and most likely with BD139/140 and 2SC5200/2SA1943 pairings. I will look at alternative drivers but this is a good as any place to start.

Trouble is that with higher rails/powers and more difficult loads we start to run out of steam with a single pair of outputs. The input impedance of the power devices drops at high current (so called beta droop) which taxes the drivers. In the CFP the drivers feed the base of the output devices with their collectors - a high impedance drive. This is the worse case for driving power BJTs at high currents with beta-droop. Even the venerable 2SC5200's have some fall off in beta at high current. The other issue is BJTs are easily destroyed if over-stressed. The solution is usually to add another pair of outputs.

There have been CFP output designs with more than one pair of outputs, some use one driver per pair of power devices, others have a separate dedicated driver per power device so a pair of outputs needs two drivers. I remember Sakis saying that none were quite as successful as the single output pair. So I'm keeping to the single output pair.

However - by adding a pair of power MOSFETs that only turn-on when the power BJTs reach high current we can divert the extra load current away from the BJTs and cap the maximum current flow through them. Under difficult load conditions or peak powers the MOSFETs will take on the additional stress and protect the BJTs. These MOSFETs are more robust and do not have beta droop. Since the MOSFETs remain off during low-medium powers they do not contribute to the sound of the output stage for normal listening (whatever that is).

So we preserve the sound of the single output pair at low-to-medium powers but enable the amplifier to drive higher peak powers into more difficult loads when needed
Attached Images
File Type: jpg temp.jpg (192.2 KB, 3008 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; 10th November 2013 at 06:56 PM.
  Reply With Quote
Old 10th November 2013, 09:58 PM   #10
AKSA is offline AKSA  Australia
diyAudio Member
 
Join Date: Sep 2001
Location: Melbourne, Australia
Gareth,
Clever approach.
The mosfets saturate much further from the rail than the bipolars. You could be in a situation where the bipolar is approaching rails, but when the mosfet cannot turn off because the gate drive is too close to the rails.
How about substituting the devices so that at lower output the mosfet operates and at high output the bipolar chimes in to help? Would this work?

Hugh
__________________
Aspen Amplifiers P/L (Australia)
www.aksaonline.com
  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
Rod Elliot P3A Layout - Critics paranoid06 Solid State 338 14th September 2014 04:29 PM
Fixed line level BSC based on Rod Elliot's circuit arjscott Multi-Way 10 8th November 2006 05:57 AM
Rod Elliot P3a boards stuffed FS! Mattyo5 Swap Meet 8 3rd October 2006 01:45 PM
Rod Elliot's P3A considirations Sander Solid State 14 25th September 2005 11:38 AM
Rod Elliot P3A amp noises tktran Solid State 2 18th March 2003 07:04 PM


New To Site? Need Help?

All times are GMT. The time now is 09:21 PM.


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