All,
I am looking for verified Audio improvement suggestion on my Leach Super Amp.
You can see picture of my am at http://community.webshots.com/user/vuamp2.
The components I used match exactly to the list recommended in the Leach Super Amp site.
I am looking for verified Audio improvement suggestion on my Leach Super Amp.
You can see picture of my am at http://community.webshots.com/user/vuamp2.
The components I used match exactly to the list recommended in the Leach Super Amp site.
armchair engineering suggestions to be sure...
The suggestions on C6 aren’t up-to-date; nonpolar Al electro is the only option at these values, series polar caps have been shown (by Cyrill Bateman) to be much worse for distortion, Cyrill also recommends larger V rating for the nonpolar caps, possibly series 2 nonpolar for best possible with electros
If you’re using these in a system with a subwoofer, you could consider raising the corner freq on the main amps to the point that metallized mylar or even polypropylene caps could fit – a DC servo or even DC coupling (with the addition of some offset trimming adj) are other options
I would consider adding a pair of resistors from the amp output to the diff pair bias Rs to get a bootstrap constant current bias rather than modulating the diff pairs with the common mode voltage and hoping symmetry cleans up the even order distortion
Overall I would say the amp has too little open loop gain, I realize that some of that is attributable to the low TIM design concept – but I think better options exist
While the V gain may be low by design (and largely through the application of local, linearizing, emitter degeneration), I really don’t think the current gain has any excuse for being so low, I would expect the linearity to be severely compromised by the nonlinear load presented to the VAS
Without addressing the low current gain, I doubt improving the input and vas cascode/cascade circuits would help much
Leach Super Amp schematic:
http://users.ece.gatech.edu/~mleach/superamp/circuit.pdf
at the Leach Super Amp page:
http://users.ece.gatech.edu/~mleach/superamp/
The suggestions on C6 aren’t up-to-date; nonpolar Al electro is the only option at these values, series polar caps have been shown (by Cyrill Bateman) to be much worse for distortion, Cyrill also recommends larger V rating for the nonpolar caps, possibly series 2 nonpolar for best possible with electros
If you’re using these in a system with a subwoofer, you could consider raising the corner freq on the main amps to the point that metallized mylar or even polypropylene caps could fit – a DC servo or even DC coupling (with the addition of some offset trimming adj) are other options
I would consider adding a pair of resistors from the amp output to the diff pair bias Rs to get a bootstrap constant current bias rather than modulating the diff pairs with the common mode voltage and hoping symmetry cleans up the even order distortion
Overall I would say the amp has too little open loop gain, I realize that some of that is attributable to the low TIM design concept – but I think better options exist
While the V gain may be low by design (and largely through the application of local, linearizing, emitter degeneration), I really don’t think the current gain has any excuse for being so low, I would expect the linearity to be severely compromised by the nonlinear load presented to the VAS
Without addressing the low current gain, I doubt improving the input and vas cascode/cascade circuits would help much
Leach Super Amp schematic:
http://users.ece.gatech.edu/~mleach/superamp/circuit.pdf
at the Leach Super Amp page:
http://users.ece.gatech.edu/~mleach/superamp/
If you have problem with it try 
Why not two of these for c6 in a perfectionist design?:
http://www.partsexpress.com/pe/pshowdetl.cfm?&PartNumber=027-447&DID=7
The voltage rating is much higher than necessary but I don't know of lower voltage types at a reasonable price, I'll keep looking.
Most higher voltage polar caps in this capacitance range self resonate around 20 kHz, and are inductive above. I've not measured the smaller electros specified in the Leach design but I'd take an educated guess in the 100s of kHz range where the amp still has enough bandwidth to cause problems:
http://www.capacitors.com/picking_capacitors/pick8lg.jpg
The design obviously works but I wonder if this was taken into consideration? This combined with the other flaws of electros is a good reason to consider these poly caps in a perfectionist design.
Thoughts? This is probably overkill?
http://www.partsexpress.com/pe/pshowdetl.cfm?&PartNumber=027-447&DID=7
The voltage rating is much higher than necessary but I don't know of lower voltage types at a reasonable price, I'll keep looking.
Most higher voltage polar caps in this capacitance range self resonate around 20 kHz, and are inductive above. I've not measured the smaller electros specified in the Leach design but I'd take an educated guess in the 100s of kHz range where the amp still has enough bandwidth to cause problems:
http://www.capacitors.com/picking_capacitors/pick8lg.jpg
The design obviously works but I wonder if this was taken into consideration? This combined with the other flaws of electros is a good reason to consider these poly caps in a perfectionist design.
Thoughts? This is probably overkill?
Hi Vuamp2,
Very nice amplifier by the way!
I've not built or worked on any of the Leach designs but have followed them from the beginning. It's been on my mind to try poly caps in the feedback location in the amplifiers that I'm currently using for several years. I'd probably look for ways to get the required value below 100 uF, probably scale the feedback network and live with a bit more DC offset.
It's interesting that the DC balance of the +/- diff amp input is only valid with AC coupled sources since there is no input blocking cap. I believe that grounding the input will upset this balance and increase the DC offset at the output.
The caps that I refer to are physically large with their much higher working voltage but I don't know of a better solution, other than servos or DC coupling.
We should ask, as in the other thread where I posed the question as to why distortion increases at HF, why does distortion increase at low frequencies even below clipping? The nonlinearities in this feedback cap are one reason.
I would not expect it to make an audible difference with speakers given the high levels of distortion that most speakers produce. Still it seems that there should be a better solution for this cap in designs such as this where there's so much attention to detail.
I've not seen anyone discuss the issue of capacitor self resonance and the HF inductive component for this feedback cap. It approximates a series resonance circuit and the .1 uF in parallel is just a very good cap in parallel with this series resonance circuit, it might actually introduce another parallel resonance but I've not measured it. The design obviously works but I wonder if this may explain part of the mismatch between simulation and real hardware, and the need for tweaking the HF feedback network.
Pete B.
Very nice amplifier by the way!
I've not built or worked on any of the Leach designs but have followed them from the beginning. It's been on my mind to try poly caps in the feedback location in the amplifiers that I'm currently using for several years. I'd probably look for ways to get the required value below 100 uF, probably scale the feedback network and live with a bit more DC offset.
It's interesting that the DC balance of the +/- diff amp input is only valid with AC coupled sources since there is no input blocking cap. I believe that grounding the input will upset this balance and increase the DC offset at the output.
The caps that I refer to are physically large with their much higher working voltage but I don't know of a better solution, other than servos or DC coupling.
We should ask, as in the other thread where I posed the question as to why distortion increases at HF, why does distortion increase at low frequencies even below clipping? The nonlinearities in this feedback cap are one reason.
I would not expect it to make an audible difference with speakers given the high levels of distortion that most speakers produce. Still it seems that there should be a better solution for this cap in designs such as this where there's so much attention to detail.
I've not seen anyone discuss the issue of capacitor self resonance and the HF inductive component for this feedback cap. It approximates a series resonance circuit and the .1 uF in parallel is just a very good cap in parallel with this series resonance circuit, it might actually introduce another parallel resonance but I've not measured it. The design obviously works but I wonder if this may explain part of the mismatch between simulation and real hardware, and the need for tweaking the HF feedback network.
Pete B.
Leach Super Amp with MJL1302A/3281A
Okay. I just completed two channels of Leach Super Amp with MJL1302A and MJL3281A output transistors. Preliminary test was good – they put out sound without hiss or distortion.
I am going to compare this amp to my original Leach Super Amp (with T03 output transistors) soon. Will there be any surprises?
I use 4 of the CPU heatsinks per channel. They seem to be a little too small???????
I want to get this basic mod to work before I try any of the suggestions I got so far. Stay tune
You can see pictures at this link
Okay. I just completed two channels of Leach Super Amp with MJL1302A and MJL3281A output transistors. Preliminary test was good – they put out sound without hiss or distortion.
I am going to compare this amp to my original Leach Super Amp (with T03 output transistors) soon. Will there be any surprises?
I use 4 of the CPU heatsinks per channel. They seem to be a little too small???????
I want to get this basic mod to work before I try any of the suggestions I got so far. Stay tune
You can see pictures at this link
PB2 said:
PB2, this could be my long term plan. Between now and the new year I just want to focus on the low hanging fruit first
.jacco vermeulen said:
Jacco, I decided to stay with wire connected because I don't believe I will see any improvements in sound quality if I go with PCB for these output transistors. It would be nice to have these Transistors solder directly to the Amp PCB but this will require a new PCB.
I just wonder if anyone has tried this?
LEACH AMP
HI VUAMP2
for a rather profound bass response for your Leach amp (120 watt low tim) do some experiments on lowering down the values of resisitors: R34,R35=180ohm, and R36=47ohm. just try to play with this
components up or down or you might only change R36 to 47ohm or lower down to 22 ohm you should hear some changes on your output.
HI VUAMP2for a rather profound bass response for your Leach amp (120 watt low tim) do some experiments on lowering down the values of resisitors: R34,R35=180ohm, and R36=47ohm. just try to play with this
components up or down or you might only change R36 to 47ohm or lower down to 22 ohm you should hear some changes on your output.
Vuamp2 said:
I've only taken a quick look at this design, but yes it looks as if they removed both the series VAS and output stage, and instead used paralleled output devices. I didn't see a max supply voltage or power rating for the larger design.
Simple parallel connection is possible today with higher voltage output devices that are now available. One should compare SOA for both connections since it is sometimes much better with the series connection.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.