Thank you @cubicincher and @ItsAllInMyHead, will seek your advise and make notes on all that were mentioned, thank you again for taking your time to list out.
I strongly advise not installing the “turbo” diodes in the amplifier circuit. (The MUR3020W in the amp schematic.)
This will greatly aid in stability, at the small cost of a few volts of swing at very high power outputs. You can build the circuit identically with no other changes and merely omit them.
This will greatly aid in stability, at the small cost of a few volts of swing at very high power outputs. You can build the circuit identically with no other changes and merely omit them.
Dear all,
I'm building a stereo F5-T based on V3.0 boards of the store. The "simplest" one: 50W, 32V rails, no cascode.
The F5 (no turbo) guides recommend to use at least 16AWG wires, so I bet that this is the same for F5-T although the boards holes are definitely too small for that (20 awg seems to be the hole size based on different posts in this thread and based on what I can see/try). What is the minimal wire you are using to build this amp? Would 18 AWG be ok? For the PSU to amp and for the frontend<->N.P-channel board link?
Many thanks.
P.S. Posts I read which gave some informations concerning wire sizes: around #21, #80/#81, #97 and #941 and also #77 from @6L6.
I'm building a stereo F5-T based on V3.0 boards of the store. The "simplest" one: 50W, 32V rails, no cascode.
The F5 (no turbo) guides recommend to use at least 16AWG wires, so I bet that this is the same for F5-T although the boards holes are definitely too small for that (20 awg seems to be the hole size based on different posts in this thread and based on what I can see/try). What is the minimal wire you are using to build this amp? Would 18 AWG be ok? For the PSU to amp and for the frontend<->N.P-channel board link?
Many thanks.
P.S. Posts I read which gave some informations concerning wire sizes: around #21, #80/#81, #97 and #941 and also #77 from @6L6.
18 AWG is more than sufficient for: a longer than normal run with an acceptable temperature under normal operating conditions. It will also allow enough voltage and current headroom for a fault condition that would blow a fuse before the wire would even get warm to the touch, IMO.
Someone could double/triple check though.
https://www.omnicalculator.com/physics/dc-wire-size
Someone could double/triple check though.
https://www.omnicalculator.com/physics/dc-wire-size
32V x 6.5mA = 208mW
(208 x 0.2) + Tambient (assuming 200C/w thermal resistance of typical TO-92 package).
If your amp runs 40C that means the JFET will be running 80C Tdie. You've to derate to half power at this temperature, and you're already exceeding it.
Plus, the J74 breaks down at 25V so you're on a wing and a prayer at 32V without cascode.
The cascodes exist for a reason. Use them.
(208 x 0.2) + Tambient (assuming 200C/w thermal resistance of typical TO-92 package).
If your amp runs 40C that means the JFET will be running 80C Tdie. You've to derate to half power at this temperature, and you're already exceeding it.
Plus, the J74 breaks down at 25V so you're on a wing and a prayer at 32V without cascode.
The cascodes exist for a reason. Use them.
Thanks for the heads up @ItsAllInMyHead.
Thanks @Sangram , I was following the first post of this thread from @6L6 where it is said:
I'm indeed building a F5-T V2 with 32V Rails. Is this advice maybe too optimistic? BTW, I guess I think I'll follow your advice as it's not worth taking the risk. Can anyone comment on this? To share knowledge at least.
Many thanks to all.
Use cascode.
Strongly consider omitting the ’turbo’ diodes.
Many years have gone by since the original guide, and the avalibility of the LS Jfets has become more common, as well as many successful builds by members, and the overall knowledge pool has increased a lot since then. It’s probably worth updating that part of the guide. 🙂
Strongly consider omitting the ’turbo’ diodes.
Many years have gone by since the original guide, and the avalibility of the LS Jfets has become more common, as well as many successful builds by members, and the overall knowledge pool has increased a lot since then. It’s probably worth updating that part of the guide. 🙂
Thanks @6L6 and double thanks @Sangram for initially taking care of this although it was not my initial question 🙂
Seems I need to take a second read at "fresh" posts.
To sum this up, I'll remove the MUR3020W, will cascode and will populate C3,C4 which I decided to ignore until now for an unknown reason.
Any recommendation concerning the clip-on heatsinks to use for the cascode transistors? I found 230-75AB at MSR for TO-220, not sure if it's ok.
Seems I was at risk - So many thanks and have a nice day all
Seems I need to take a second read at "fresh" posts.
To sum this up, I'll remove the MUR3020W, will cascode and will populate C3,C4 which I decided to ignore until now for an unknown reason.
Any recommendation concerning the clip-on heatsinks to use for the cascode transistors? I found 230-75AB at MSR for TO-220, not sure if it's ok.
Seems I was at risk - So many thanks and have a nice day all
I also use 32V rail, took out the diodes and did cascode. Set the cascode voltage in the middle of the 32V by changing resistors. Used a cliip-on heat sink for TO220 on the BJT's, but they heat only little. Really easy and works great.
Back with some bias "problem" 🙂
So, I removed Turbo diodes, added cascode so that I have half the Vrail presented to the jfet (measured around 13V) and nothing smelt strange. So far so good.
So the configuration is: F5-T v2, ~32V rails, cascode (10k/10k), no turbo diode, jfet quad 6-8mA from the diyaudiostore and all other transistors also sourced from the diyaudiostore.
The max bias I managed to get on both Left and right are:
157mV for P-Channel, 164mV for N-Channel with 0mv at out on one hand
and 159mV for P-Channel, 169mV for N-Channel with 0mv at out for the other.
As far as I understand it may be related to too low Idss of the JFET(?) and may be fixed by increasing R5 and R6 (ref from the V3 schematic) from 1k to 2k or 3k.
BTW, I prefer to ask as my bias seems quite low compared to others who had the same kind of issue and I don't want to rush if it may be a symptom of something else that I may have missed.
Any advice welcome here 🙂
Many thanks.
So, I removed Turbo diodes, added cascode so that I have half the Vrail presented to the jfet (measured around 13V) and nothing smelt strange. So far so good.
So the configuration is: F5-T v2, ~32V rails, cascode (10k/10k), no turbo diode, jfet quad 6-8mA from the diyaudiostore and all other transistors also sourced from the diyaudiostore.
The max bias I managed to get on both Left and right are:
157mV for P-Channel, 164mV for N-Channel with 0mv at out on one hand
and 159mV for P-Channel, 169mV for N-Channel with 0mv at out for the other.
As far as I understand it may be related to too low Idss of the JFET(?) and may be fixed by increasing R5 and R6 (ref from the V3 schematic) from 1k to 2k or 3k.
BTW, I prefer to ask as my bias seems quite low compared to others who had the same kind of issue and I don't want to rush if it may be a symptom of something else that I may have missed.
Any advice welcome here 🙂
Many thanks.
nothing stops you from increasing these resistors
back down trimpots/Iq prior to change
back down trimpots/Iq prior to change
Replaced R5 and R6 with 2.2k that I had and finally biased it. It sounds incredibly well. It's absolutely gorgeous. Thank you so much for your support, I learnt a lot and you know what? I'm simply happy thanks to all of you ;-)
Congratulations, blacknoz! I have mine now up and running as well. Improved a lot on the hum. In the end the small hum that I had left came from the input signal lines. They go from the RCA cores to the amps on both sides, and with the ground in the PSU connections to each other. Solved it by having one signal line go straight from RCA receptacle to the RHS amp, and the other input line follows the same path to the RHS amp, crosses to the other LHS amp closely bound to the PSU-lines. All hum has gone away. 50 Hz is gone now, 100 Hz and higher 50 Hz harmonics are around 100 dB below the 1 kHz test tone. H2 is rather high: -60 dB. But I did some simulations in LTSpice, and found out that with the IRFP mosfets one go a little lower, but only down to -80 dB. That will be my next challenge.