Hi Andrew, have you actually done these tests and can say that based on your listening experiences which is your preferred method?
Can you point me to what methods for this particular arrangement. I have my own pcb, in the works, so could use some suggestions as to design improvement. I know for #1, a bigger HS
Stammheim, based on AndrewT tests?/suggestions, I suggest to start with the SMPS connected to both LME and FETs and then go from there, change, if absolutely necessary, to gain a few extra watts or clipping behavior.
Rick
As I see the matter,
in my opinion and experience, about having different power supplies for the input stage and output stage, the advantages are:
It does not inject the output stage ripple to the input stage.
And; it reduces the interchannel interference.
And:
As the PSRR of the LME is actually not that good in higher frequencies, it is easier to get better results by having an old fashion linear regulated supply for the LME.
SMPS is OK for the output stage, still well filtered though...
Should the input stage have higher voltage than the output stage?
I do not have an opinion about that. It may be a design decision.
The LME is however not capable of rail to rail output. It swings only to about 4 volts down from its rails. How linear it is near its maximum swing, I do not know.
Should it therefore have a bit higher rail voltages, compared to its output stage, for the optimum performance? A decision you must make...
in my opinion and experience, about having different power supplies for the input stage and output stage, the advantages are:
It does not inject the output stage ripple to the input stage.
And; it reduces the interchannel interference.
And:
As the PSRR of the LME is actually not that good in higher frequencies, it is easier to get better results by having an old fashion linear regulated supply for the LME.
SMPS is OK for the output stage, still well filtered though...
Should the input stage have higher voltage than the output stage?
I do not have an opinion about that. It may be a design decision.
The LME is however not capable of rail to rail output. It swings only to about 4 volts down from its rails. How linear it is near its maximum swing, I do not know.
Should it therefore have a bit higher rail voltages, compared to its output stage, for the optimum performance? A decision you must make...
I would like to use higher quality input capacitors like Mundorf 2.2uf so in balanced configuration I see we have 330uf and we cant get the higher quality film caps in that case so how do we reduce the value to like 2.2uf? so what resistors needs to be changed? can we use 2.2uf in series with 1.2k resistor at the input?
This really depends on your application. Using a lower value input cap with a low input impedance results in low frequency rolloff. In the case of a 2.2uF cap and 1.2k differential input impedance, you're looking at:
Fc = 1/(2*Pi*R*C) where R and C are the respective input impedance and cap.
This gives a 60Hz -3dB point which would be fine if you're using a subwoofer and only powering satellites with the amp, but it would be a disaster if you are powering full range speakers.
In order to get the rolloff to an appropriate minimum level you would need 10 times the input impedance, so around 12k.
If you do this, your feedback resistors also need to be increased by a factor of 10 to maintain the correct gain. These increased resistor values cause the amplifier to have a significantly higher noise floor.
The noise floor difference will be audible, as for the caps, I'm not so sure. I have carefully measured the organic polymer caps chosen on the input, and they work very well as coupling caps. There was no measurable difference between them and a good pair of film caps.
Before embarking on the above changes, do you really even need the input coupling caps? You can omit them entirely as long as your source does not have DC at the output. Generally speaking, any properly built source should not suffer that problem.
Regards,
Owen
I have a question regarding the PSU for this amp: You recommend 2 separate supplies, one for the IC and one for the FETs, and you mention a regulated supply for the IC being advantageous. How would I go about making one of those? As far as I can tell from browsing Mouser, there are no linear regulator ICs that will do 50V+, which is what I'm looking for on the LME, and I'm very skeptical of switched-mode supplies - would it be worth the trouble of making (and maybe even designing) an SMPS instead of a good ol' unregulated supply with a bunch of really big caps thrown in? Or, even better, can you recommend a linearly regulated supply design that will do 45-50V, if such a thing even exists?
Thanks,
George
Thanks,
George
Last edited:
Boards are still available here,
http://www.diyaudio.com/forums/vend...jects-available-here-bal-bal-se-se-lpuhp.html
you will have to source your own parts
http://www.diyaudio.com/forums/vend...jects-available-here-bal-bal-se-se-lpuhp.html
you will have to source your own parts
bitizzle, you can use the popular lm317/lm337 to make a regulator for the front end. As long as the input voltage - the regulated output voltage is less than 45 you are fine. ie I used the 317/337 to make +/- 75volt regulators for the LME chips with +/- 85volt input. I used +/- 68 volts for the fets, check the application notes for the lm317/337. If your power transformer does not have a higher secondary voltage, you can used a low power 1-5VA transformer with dual 10 volt or so secondaries to add to your power transformer.
I had a bit of a problem with my version of the amplifier. After working flawlessly for over a year of daily use I noticed a slight burning smell. Not circuit board burning smell but more like an electric heater that hasn't been used in a couple of years smell. The amplifier chasis was way too hot to touch over the left channel. I would get this smell with the amplifier on and no music playing. The amplfier went on the work bench and I found the bias current on the left channel jumping all over the place from 80 ma to 450ma. On my build the heatsinks are a bit small so I chose a setting of 150 ma. With a setting of 150 ma the amplifier would dissapate about 20 watts per channel and would get a bit warm. A small 1/32 turn of the bias pot fixed it and the bias is now stable again at 145ma. The pot is obviously bad. I will replace the pot with a resistor in the near future. I will keep using the amplifier as is until I can make some sort of chasis for the LPUHP boards. It take about 30-45 minutes to dissasemble the amplifier so I can change the pot for a resistor.
Yikes! Not good...
Did everything else survive?
I suppose it certainly is best to swap the pots for actual resistors once you've found the level, but I've never personally experienced this issue.
Alternatively, the resistors around the pot could possibly be adjusted to provide a "fail safe" level, but this will vary depending on the implementation.
Let me know how the fix goes... hopefully everything else is alright.
Are you using the pots from the BOM? If not, what make/type/value are they?
Cheers,
Owen
Did everything else survive?
I suppose it certainly is best to swap the pots for actual resistors once you've found the level, but I've never personally experienced this issue.
Alternatively, the resistors around the pot could possibly be adjusted to provide a "fail safe" level, but this will vary depending on the implementation.
Let me know how the fix goes... hopefully everything else is alright.
Are you using the pots from the BOM? If not, what make/type/value are they?
Cheers,
Owen
I have a schematic drawn that should:
Allow input from SMPS to switch from low bias current during SMPS start up then some delay with 555 timer before switching to 'full bias' current that is based on current heat sink temperature.
Have two operating bias settings based on temperature (summer/winter or similar) using a daughter PCB that mounts in the position of the pot that has 3 resistors and 2 smd relays.
Overtemp shutdown
Momentary switch to control SMPS on/off or SSR power control.
Momentary switch for mute
A few other things I've forgotten.
When I looked at it last (12months ago or something probably) I had trouble working out a PCB layout it was the first layout I'd ever looked at seriously and there was a lot of interconnected devices trying to work on a 2 layer PCB and it was too much for me at the time. I could revive it if it would be useful for a few 'The Wire' builders? I'd want to do some thorough testing of it before sending out PCBs though. At the moment it's just TTL logic and comparators. I could convert it to be a mix of logic ICs and an arduino that would make the PCB design simpler.
Thoughts?
Cheers,
Chris
Allow input from SMPS to switch from low bias current during SMPS start up then some delay with 555 timer before switching to 'full bias' current that is based on current heat sink temperature.
Have two operating bias settings based on temperature (summer/winter or similar) using a daughter PCB that mounts in the position of the pot that has 3 resistors and 2 smd relays.
Overtemp shutdown
Momentary switch to control SMPS on/off or SSR power control.
Momentary switch for mute
A few other things I've forgotten.
When I looked at it last (12months ago or something probably) I had trouble working out a PCB layout it was the first layout I'd ever looked at seriously and there was a lot of interconnected devices trying to work on a 2 layer PCB and it was too much for me at the time. I could revive it if it would be useful for a few 'The Wire' builders? I'd want to do some thorough testing of it before sending out PCBs though. At the moment it's just TTL logic and comparators. I could convert it to be a mix of logic ICs and an arduino that would make the PCB design simpler.
Thoughts?
Cheers,
Chris
Very good idea. Easy to implement and a decent safeguard.
Any suggestions that don't involve a relay?
A few Recent Threads on using fast acting solid state relays assembled from appropriate devices that do not interfere with sound quality.
Some locations offer different types of protection.
In the SPKR output
In the SPKR Return
In the Supply rails
Across the speaker to audio ground.
All of these can be "add ons" not on the amp PCB.
But many would suit smd on the PCB and still retain the compactness you have achieved.
A suggestion if you do design an "all protected PCB" :
add space for more air circulation around the driver heatsink, including getting the capacitors away from the hot surfaces.
Or even better: a bigger heatsink for the higher supply voltage Builders.
Some locations offer different types of protection.
In the SPKR output
In the SPKR Return
In the Supply rails
Across the speaker to audio ground.
All of these can be "add ons" not on the amp PCB.
But many would suit smd on the PCB and still retain the compactness you have achieved.
A suggestion if you do design an "all protected PCB" :
add space for more air circulation around the driver heatsink, including getting the capacitors away from the hot surfaces.
Or even better: a bigger heatsink for the higher supply voltage Builders.
Last edited:
Why not control the bias with an optocoupler?
Short one resistor from the bias circuit with that to allow a lower bias setting?
Makes the bias circuit nice and compact compared to relays and stuff?
Could also make the bias adjustable by varying the opto led brightness???
Nothing else appeares to be damaged. Once the bias current was set to 145ma, I left it powered on for > 24hours and checked every few hours and it was steady. The amplifier is back in the living room and lets say I shook some of the dust off the speakers with Led Zepplin, Hendrix and Emmerson Lake and Palmer. Everything appears to be fine.
Yes the pot was from the BOM.
There is speaker relay DC protection.
There are fuses for the amplifier boards and a line fuse if things go real south.
I must have adjusted thousands of multi turn trim pots over the years and I don't recall ever having to do a board replacement because of a multi turn trim pot going bad. Single turn pots, yes
Yes the pot was from the BOM.
There is speaker relay DC protection.
There are fuses for the amplifier boards and a line fuse if things go real south.
I must have adjusted thousands of multi turn trim pots over the years and I don't recall ever having to do a board replacement because of a multi turn trim pot going bad. Single turn pots, yes
Hi all, I'm reading this interesting thread.
I already have a large number of pieces, 2 transfos (200VA/35V), 2 unreg power supplies (≈+/-50V), source selection board, soft-start, heat-sinks ...
I wanted to build 2 SKA amps in dual mono, but finally I think I'll go with The Wire.
1) I add myself to the GB list and OPC contact me for the payment?
I'd like to see more wiring diagrams, rookie inside, 1st amp build
I also have a question, my relay based source selection board is designed in dual mono, no problem, but I would add a volume control Dantimax RelVol1C RelVol1 & 3
2) Can I keep my separate signal masses (dual mono) with this volume board control?
3) For the attenuator resistors, 20K value is ok?
Thanks a lot guys
I already have a large number of pieces, 2 transfos (200VA/35V), 2 unreg power supplies (≈+/-50V), source selection board, soft-start, heat-sinks ...
I wanted to build 2 SKA amps in dual mono, but finally I think I'll go with The Wire.
1) I add myself to the GB list and OPC contact me for the payment?
I'd like to see more wiring diagrams, rookie inside, 1st amp build
I also have a question, my relay based source selection board is designed in dual mono, no problem, but I would add a volume control Dantimax RelVol1C RelVol1 & 3
2) Can I keep my separate signal masses (dual mono) with this volume board control?
3) For the attenuator resistors, 20K value is ok?
Thanks a lot guys