Re: Re: Re: Re: SC ULD 200W AB Amp
Mihai,
Thank you for the explanation.
A bit OT.
I’m puzzled by the 7.5mA bias current. The original FC-100 schematic shows 170mA Iq and 2mA through the bias network. Where does 7.5mA come from? Has something changed?
Anyway, impressive (to me) Delta 2Vre at 1.5mV.
Brian.
roender said:
Mihai,
Thank you for the explanation.
A bit OT.
I’m puzzled by the 7.5mA bias current. The original FC-100 schematic shows 170mA Iq and 2mA through the bias network. Where does 7.5mA come from? Has something changed?
Anyway, impressive (to me) Delta 2Vre at 1.5mV.
Brian.
SandyK
If I properly understood your question, the PCB is double sided. It was laid out using Sprint Layout. The shot that I posted just showed the top side. Sprint has a free viewer should anyone want to see view both layers and check my work. I have no problem giving away the '.LAY' file
http://www.abacom-online.de/uk/html/sprint-layout.html
If I properly understood your question, the PCB is double sided. It was laid out using Sprint Layout. The shot that I posted just showed the top side. Sprint has a free viewer should anyone want to see view both layers and check my work. I have no problem giving away the '.LAY' file
http://www.abacom-online.de/uk/html/sprint-layout.html
Sc Uld2
Carl
I was mainly pointing out that L and R channels use mirror
image PCBs to permit a far tidier internal layout.
SandyK
P.S.
The additional 130 ohm resistor shown, would be best replaced by a 200 or 500 ohm multi turn trimpot. The 130 ohm was only a guesstimate based on the figures from the original design, and made simple to encourage some people to try it, as Alfred Rofe did,and reported back. Fine tuning this value will further improve sound quality if the LTP devices are well matched for VBE and HFE.
Carl
I was mainly pointing out that L and R channels use mirror
image PCBs to permit a far tidier internal layout.
SandyK
P.S.
The additional 130 ohm resistor shown, would be best replaced by a 200 or 500 ohm multi turn trimpot. The 130 ohm was only a guesstimate based on the figures from the original design, and made simple to encourage some people to try it, as Alfred Rofe did,and reported back. Fine tuning this value will further improve sound quality if the LTP devices are well matched for VBE and HFE.
What I've posted is as far as I've taken it. I added the traces to pickup the internal diodes and a series resistor but no adjustable Vbe. I also made the non Vbe changes that are described in the article and this thread. In my layout the PCB is parallel to the heatsink rather than coming off at right angle as it seemed more space efficient. The output devices are mounted on the heatsink under the PCB.
I haven't built this board. I don't know if there are errors. It definitely needs to be checked with 'fresh eyes' before anyone spends the time and money to build it.
I haven't built this board. I don't know if there are errors. It definitely needs to be checked with 'fresh eyes' before anyone spends the time and money to build it.
Nice layout Carl, tres purdy......
Nigel,
My choice would be for a Vbe multiplier, but this would change the output devices of course, greatly altering the pcb. If using five internal diodes for bias control, the current through the string would need to be reduced to Mihai's recommended 2-3mA for correct bias levels. This might require a CFP driver on both sides, with the master device operating at about 3mA - and a higher level of complexity. However, bearing in mind the excellence of Mihai's FC-100 amp, this might be well worth it.
Sandy,
Provision could be easily made in a new pcb for a resistor to ensure equal Vce on the LTP devices, and for a diode to be inserted into the CB link on the inactive CM device. I like Carl's layout, it's much cleaner than the original. I find it's best to have pcbs for both channels identical to minimise assembly errors (idiot factor), and accommodate a LR inversion on the power supply pcb which holds the filter caps, rectifiers and bleeders.
Hugh
Nigel,
My choice would be for a Vbe multiplier, but this would change the output devices of course, greatly altering the pcb. If using five internal diodes for bias control, the current through the string would need to be reduced to Mihai's recommended 2-3mA for correct bias levels. This might require a CFP driver on both sides, with the master device operating at about 3mA - and a higher level of complexity. However, bearing in mind the excellence of Mihai's FC-100 amp, this might be well worth it.
Sandy,
Provision could be easily made in a new pcb for a resistor to ensure equal Vce on the LTP devices, and for a diode to be inserted into the CB link on the inactive CM device. I like Carl's layout, it's much cleaner than the original. I find it's best to have pcbs for both channels identical to minimise assembly errors (idiot factor), and accommodate a LR inversion on the power supply pcb which holds the filter caps, rectifiers and bleeders.
Hugh
Hello Carl,
This is probably a solution for all those who are trying to work with something other than what they have got, ie a kit
I think it is different enough not to breach copuright, however I would like to make the following comments.
Three output devices follows the origional Application note, I is my humble opion that if you should desire the expense of extra O/P devices you may wish to use this amp as a PA, surely the DIY quest is for quality rather than quantity an the volume that we play at at home will never take advantage of the full comliment of O/P devices, in the direction of overkill, the application note may use three pair of devices but they are in the business of selling components.
There is a large space between the rear of the O/p transistors and the possible location of the heatsink, this should be as close to the edge as possible to allow the trnasistors to be mounted in the middle of the heatsink and utilise the heatsink to its full potential, also the O/P transistors could be spaced further apart to this same end, if we have revolution lets make sure we intelligently apply all posibilities to the best outcome.
On the quality side you have not made provision for SandyK's trimpot on the LTP, this has been the consistant basis of my input to this thread, because of it obvious achievemnt in Sound Quality.
As I can not see all the tracks, but it looks like you have put in some of the mods suggested, one extra thing you may put in is an extra pin on Q8 to allow changing the transistor to a different configuration without have to cross its legs
But all in all a bold and couragous start
Alfred
This is probably a solution for all those who are trying to work with something other than what they have got, ie a kit
I think it is different enough not to breach copuright, however I would like to make the following comments.
Three output devices follows the origional Application note, I is my humble opion that if you should desire the expense of extra O/P devices you may wish to use this amp as a PA, surely the DIY quest is for quality rather than quantity an the volume that we play at at home will never take advantage of the full comliment of O/P devices, in the direction of overkill, the application note may use three pair of devices but they are in the business of selling components.
There is a large space between the rear of the O/p transistors and the possible location of the heatsink, this should be as close to the edge as possible to allow the trnasistors to be mounted in the middle of the heatsink and utilise the heatsink to its full potential, also the O/P transistors could be spaced further apart to this same end, if we have revolution lets make sure we intelligently apply all posibilities to the best outcome.
On the quality side you have not made provision for SandyK's trimpot on the LTP, this has been the consistant basis of my input to this thread, because of it obvious achievemnt in Sound Quality.
As I can not see all the tracks, but it looks like you have put in some of the mods suggested, one extra thing you may put in is an extra pin on Q8 to allow changing the transistor to a different configuration without have to cross its legs
But all in all a bold and couragous start
Alfred
SC ULD2
"Three output devices follows the origional Application note, I is my humble opion that if you should desire the expense of extra O/P devices you may wish to use this amp as a PA, surely the DIY quest is for quality rather than quantity an the volume that we play at at home will never take advantage of the full comliment of O/P devices, in the direction of overkill, the application note may use three pair of devices but they are in the business of selling components."
Alfred
Cut to the chase, why don't you ?
Perhaps Mihai is in the business of selling his own amplifier designs too?
SandyK
"Three output devices follows the origional Application note, I is my humble opion that if you should desire the expense of extra O/P devices you may wish to use this amp as a PA, surely the DIY quest is for quality rather than quantity an the volume that we play at at home will never take advantage of the full comliment of O/P devices, in the direction of overkill, the application note may use three pair of devices but they are in the business of selling components."
Alfred
Cut to the chase, why don't you ?
Perhaps Mihai is in the business of selling his own amplifier designs too?
SandyK
Sandy,
I don't believe that's true. Mihai offered his design gratis into the public domain, and makes a fine living as a professional in IT, why would he want to profit from it?
I feel that for domestic high end this design is best implemented with a simple BD139 Vbe multiplier, and that C5200/A1943 devices are most appropriate, and certainly much cheaper. I feel that 0.33R emitter degeneration is required for best reliability.
Sandy's suggestions about current mirrors and balancing the LTP are great ideas, and here's a few more:
3. Silver mica lag compensation.
4. Silver mica phase lead between collector of VAS and fb node.
5. 2SC3423 VAS running at around 10mA, with 10R emitter degeneration, and well heatsinked.
6. No EF to drive the VAS, direct connection from input side of LTP to VAS base.
7. Charge suckout across the driver emitters.
This would make a very nice amplifier with very low distortion, great imaging, and a lovely, airy sound.
Cheers,
Hugh
SC ULD2
Hugh
I apologise if I am in error. what I meant to say was:
"Perhaps Mihai is in the business of selling his own design amplifiers too ?"
Perhaps it is the passion that he comes across with, in his reasons why other designers' amplifiers have such glaring design inadequacies, that gave me that impression ?
I am not just referring to the SC ULD2 either.
SandyK
Hugh
I apologise if I am in error. what I meant to say was:
"Perhaps Mihai is in the business of selling his own design amplifiers too ?"
Perhaps it is the passion that he comes across with, in his reasons why other designers' amplifiers have such glaring design inadequacies, that gave me that impression ?
I am not just referring to the SC ULD2 either.
SandyK
Carl_Huff said:Hugh,
I think your suggestions are certainly worthy of the effort. I am not particularly attached to the layout that I worked up. If someone wanted to jump in and integrate them into the design, personally I'd be tickled and motivated to build a couple of channels!
With a half decent schematic/pcb software package you would be able to update the schematic then forward annotate it to the PCB.
I would offer to do it with mine but dont have the time.....
That sounds really interesting. Are you able to provide (once the design has been reviewed) a file suitable for sending off to a PCB bureau, or which I could use to etch at home?Carl_Huff said:
My understanding of copyright law, as it applies to circuit layouts (at least in the US), is that only exact copies (i.e. 'facsimiles') are covered. If you recreate a layout yourself (by hand, or using software), it's not a violation of copywrite, since it has slight changes (e.g. track width, color, etc). If the circuit is implemented in silicon however, as an IC, it gets special covereage under the Semiconductor Chip Protection Act of 1984, as long as it's correctly registered, and is somewhat novel. That doesn't apply here.
PS: This is just my understanding of the legal side - not the ethical side!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.