Member
Joined 2009
Paid Member
Member
Joined 2009
Paid Member
found a few extra boards - now all up to date
looking for a suitable power transformer and aren't looking to extract full power, then try the Antek AS-2230. It gives a nice +/-42Vdc which is where I prefer to operate. The design is good to +/-50Vdc (read the thread).
1/ 'maust' - paid for 1 pair
2/ 'maust' - paid for 2nd pair
3/ 'Selwe' - paid for 1 pair
4/ 'Selwe' - paid for 2nd pair
5/ 'Selwe' - paid for 3rd pair
6/ ‘bvtrinh’ - paid for 1 pair
7/ 'bullittstang' - paid for 1 pair
8/ 'bullittstang' - paid for 2nd pair
9/ still available
10/ still available
11/ still available
12/ still available
looking for a suitable power transformer and aren't looking to extract full power, then try the Antek AS-2230. It gives a nice +/-42Vdc which is where I prefer to operate. The design is good to +/-50Vdc (read the thread).
1/ 'maust' - paid for 1 pair
2/ 'maust' - paid for 2nd pair
3/ 'Selwe' - paid for 1 pair
4/ 'Selwe' - paid for 2nd pair
5/ 'Selwe' - paid for 3rd pair
6/ ‘bvtrinh’ - paid for 1 pair
7/ 'bullittstang' - paid for 1 pair
8/ 'bullittstang' - paid for 2nd pair
9/ still available
10/ still available
11/ still available
12/ still available
I took Gareth's latest BOM and uploaded to digikey and verified all the part numbers and locations - correctly a few quantities and part #'s to get all items to price out. Hoping the link below will work, having trouble changing the BOM to public at digikey. Current pricing is `$65/channel at today's pricing in USD.
https://www.digikey.com/BOM/Create/CreateSharedBom?bomId=7792321
Please double check me, but I went line by line to verify values and quantities.
https://www.digikey.com/BOM/Create/CreateSharedBom?bomId=7792321
Please double check me, but I went line by line to verify values and quantities.
Member
Joined 2009
Paid Member
To calculate a rough wattage just look at the power dissipation in an idea 8Ohm load.
Power = current x voltage = voltage x voltage / resistance
Power (RMS for sine wave) = voltage x voltage / 2 * resistance
So, with 37V rails let's assume the output can swing within 1V, so we have +/=36V and the power becomes 36 x 36 / 2 x 8 = 81W. That seems plenty enough power to me. Do read through the thread - I remember that using lower voltage rails may require a resistor value to be adjusted in order to ensure that the dc offset adjustment circuit has the correct range to get the output to zero.
update:
1/ 'maust' - paid for 1 pair
2/ 'maust' - paid for 2nd pair
3/ 'Selwe' - paid for 1 pair
4/ 'Selwe' - paid for 2nd pair
5/ 'Selwe' - paid for 3rd pair
6/ ‘bvtrinh’ - paid for 1 pair
7/ 'bullittstang' - paid for 1 pair
8/ 'bullittstang' - paid for 2nd pair
9/ Rhinobirds - paid for 1 pair
10/ Rhinobirds - paid for 2nd pair
11/ still available
12/ still available
I have not seen oscillation in spice when BD139/40 is used. I do get isolation when MJE15034/5 is used in spice and it takes 390p + to tame it.
The issue of instability in the driver on the negative rail is discussed in the ESP forums, but you have to dig a bit to find the thread. The problem is specific to the negative rail as far as I recall. I don't think there is any benefit to adding the equivalent feedback loop to the positive rail driver.
I suspect the cap was added to accommodate the use of MJE15035/MJE15034 devices. It may not be necessary if BD139/BD140 devices are used. Maybe someone with more experience in CFP output stages can chime in.
The issue of instability in the driver on the negative rail is discussed in the ESP forums, but you have to dig a bit to find the thread. The problem is specific to the negative rail as far as I recall. I don't think there is any benefit to adding the equivalent feedback loop to the positive rail driver.
I suspect the cap was added to accommodate the use of MJE15035/MJE15034 devices. It may not be necessary if BD139/BD140 devices are used. Maybe someone with more experience in CFP output stages can chime in.
Member
Joined 2009
Paid Member
Brian, I'm not sure which schematic you are referring to. The one for TGM8 attached to post 604 can't be it.
Compensation:
The TGM8 has two-pole compensation around the compound VAS so there are two caps (100p and 22p) in series with a resistor from their junction to ground. The pcb allows for the traditional Cdom compensation with only 1 cap but that's not what I used. Rod, in his P3a, has a single Cdom (100pF) and a single stability cap on the driver for the lower CFP only as he found that was sufficient.
CFP Stability:
In designing the pcb I provided space for stability caps on both drivers - why not, just in case. In the end, I installed them both. I'm not suing a simple CFP like the P3a, but a CFP with both Bipolar and MOSFET power devices in the feedback loop of the driver. The additional gate capacitance of the MOSFET affects the switching performance of the CFP and I had to up the driver current. I installed stability caps on both drivers to 'safe rather than sorry'.
Compensation:
The TGM8 has two-pole compensation around the compound VAS so there are two caps (100p and 22p) in series with a resistor from their junction to ground. The pcb allows for the traditional Cdom compensation with only 1 cap but that's not what I used. Rod, in his P3a, has a single Cdom (100pF) and a single stability cap on the driver for the lower CFP only as he found that was sufficient.
CFP Stability:
In designing the pcb I provided space for stability caps on both drivers - why not, just in case. In the end, I installed them both. I'm not suing a simple CFP like the P3a, but a CFP with both Bipolar and MOSFET power devices in the feedback loop of the driver. The additional gate capacitance of the MOSFET affects the switching performance of the CFP and I had to up the driver current. I installed stability caps on both drivers to 'safe rather than sorry'.
Last edited:
Member
Joined 2009
Paid Member
Hang on - there are two Brian's - I was first replying to traderbam!
I should clarify - for questions about TGM8, I've added to the confusion too since the schematic in post 604 doesn't show the exact devices used (probably because the pcb layout software had a default part I wanted to use) and you need to go to the BOM to get the actual device types.
I should clarify - for questions about TGM8, I've added to the confusion too since the schematic in post 604 doesn't show the exact devices used (probably because the pcb layout software had a default part I wanted to use) and you need to go to the BOM to get the actual device types.
If I recall correctly from an earlier editions of circuit discussion published by Rod Elliott on his ESP site, the corresponding C6 cap was mentioned briefly in the earlier P3 project. Originally the value was 220pF and was necessary across the BD140 for stability but did nothing useful across the BD139. Its a similar story on the negative side driver with P3a and most simple CFP designs I've come across, regardless of the output transistor type.
Bigun, there are many schematics posted. You did put a note on first post that the as-built schematic is on page #604 but on the first post the schematic is the simple one. Which one that has been built by the builder making the review on first post?
This is a long thread, I'm still reading page 7
This is a long thread, I'm still reading page 7
Member
Joined 2009
Paid Member
I didn't observe any stability issues. I can't remember everything but I'm pretty confident that I installed both stability caps as the 'safe' route to avoid having to go back and add one later once the board was fully populated and cleaned of flux. My design is quite different from the original P3a, the CFP driver has a lot more capacitance to driver (gate of MOSFET).
The zobel on the MOSFETs has some history on my TM7 where I experienced oscillations at RF, mostly due to poor pcb layout but those FETs love to oscillate so that made me extra cautious for TGM8.
Post 604 is my as-built version with regards the circuit but you should refer to the posted BOM for the correct description of the semiconductor parts that I used. I'll post some further-cleaned up information at some point to make the schematic part descriptions consistent with the BOM.
Another board set in the mail: note that I use some left over christmas cards to mail the boards because they provide a stiff protective support - there's not hidden meaning by the use of them otherwise.
1/ 'maust' - paid for 1 pair
2/ 'maust' - paid for 2nd pair
3/ 'Selwe' - paid for 1 pair
4/ 'Selwe' - paid for 2nd pair
5/ 'Selwe' - paid for 3rd pair
6/ ‘bvtrinh’ - paid for 1 pair
7/ 'bullittstang' - paid for 1 pair
8/ 'bullittstang' - paid for 2nd pair
9/ 'Rhinobirds' - paid for 1 pair
10/ 'Rhinobirds' - paid for 2nd pair
11/ 'bsd04' - paid for 1 pair
12/ still available
Member
Joined 2009
Paid Member
Member
Joined 2009
Paid Member
hmmmm, I don't think I've ever heard the TGM8 in stereo. I listened to one channel when I first built it but I've not yet got around to boxing up two channels as a finished project.
Last night I found some metal bits and drilled some holes, tapped some holes and as a result I created a back panel to mount the mains power connector, speaker connectors and signal input connectors. I need to fix up some more bits and pieces but I'm going to challenge myself to have a finished stereo this week.
Last night I found some metal bits and drilled some holes, tapped some holes and as a result I created a back panel to mount the mains power connector, speaker connectors and signal input connectors. I need to fix up some more bits and pieces but I'm going to challenge myself to have a finished stereo this week.
Hi
I am going to reuse the power supply of my old amp, it has higher than 42Vdc rails.
Which transistors are recommended for VAS and driver stage considering the KSA1381 and its complement are unobtainable? An older BOM contains MJE340/350 and there have been references to MJE15030/31, MJE15034/35 as well. Looking for info landed straight to diyaudio.com no less: https://www.diyaudio.com/forums/solid-state/3853-bf469-470-471-472-vs-mje340-350-a-3-print.html
I don't know how credible the info and sources are, but seems MJE340/350 aren't the best.
I can get 340/350 and 15030/31 variants. Are they straight swap except for different pinout? Has anybody built the TGM8 with MJE340/350? Any other hints or considerations for high voltage variant?
I am going to reuse the power supply of my old amp, it has higher than 42Vdc rails.
Which transistors are recommended for VAS and driver stage considering the KSA1381 and its complement are unobtainable? An older BOM contains MJE340/350 and there have been references to MJE15030/31, MJE15034/35 as well. Looking for info landed straight to diyaudio.com no less: https://www.diyaudio.com/forums/solid-state/3853-bf469-470-471-472-vs-mje340-350-a-3-print.html
I don't know how credible the info and sources are, but seems MJE340/350 aren't the best.
I can get 340/350 and 15030/31 variants. Are they straight swap except for different pinout? Has anybody built the TGM8 with MJE340/350? Any other hints or considerations for high voltage variant?
That sucks.
I've now reacquainted myself with the schematic in post #604 and it is nothing like the first one. Now I remember how you current shunted the bipolars with FETs...poser.