PCB run
Hi,
I'm going to get 2 of the attached boards made for myself. I'll put the order in by the end of the week. No silkscreen or solder mask to keep cost down. If interested let me know. If not, no big deal... will be a very small run
Added degeneration resistor on J74, and extra holes in case of caddock resistors on FET source.
Hi,
I'm going to get 2 of the attached boards made for myself. I'll put the order in by the end of the week. No silkscreen or solder mask to keep cost down. If interested let me know. If not, no big deal... will be a very small run
Added degeneration resistor on J74, and extra holes in case of caddock resistors on FET source.
Last edited:
I would suggest you NOT feed a regulator, or a fan, off one of the power rails and amp gnd point - who knows, it mightn't create problems but .......
Georg went thru a similar 'trimming' process when developing his pcbs, so you might take another look at his method of doing the ground connections - altho some of the other tracking ideas may seem too 'fussy', they leave the options of doing a number of different things with the same pcb.
The type of resistor used for R13 degen next to the jfet will be quite noticeable to the sound quality - there's room enough to add donuts for a 'standard' resistor for the option
I would do the same thing for C3,4 & 5.
On the pcb design pckage, I would increase the donut size a lot to withstand various component changes - possibly increase the track widths also, particularly the high current ones.
There has been some discussion about adding gate stoppers to the input jfets - some people find them useful, others say it's wrong - again, options to find out for yourself ...
s there's room for more options for the power resistors on the Fets, suggest adding extra donuts here also - there are many 'standard' power resistors available and also, 2 in parallel work quite well - there's plenty of room ...
I would suggest another look at the grnd tracking system - perhaps Andrew could add some words here about this and possible grnd plane options
I would separate the Gnd point for the 68uF supply caps from the signal gnd point but not sure where you want to place the amp's central gnd point for chassis earth (or thermistor float?) switcher supplies, speaker returns, signal gnd, etc
Curiously, there are many pcbs with amp cct gnd, signal 0v, speaker returns, etc connected at one point and it works just fine, so this 'separate supply bypass gnd' is just another option.
Oh, just a small thing - suggest adding an extra donut for the P1, 2 as some come with pins 'in line', not 'offset' - an easy option.
If you wanted to, you could reverse the connections to R1 and the input terminal could be arranged for a socket or terminal block with the input gnd ...
I hope some of this could be of some use to you - just suggestions, mind you ....
Georg went thru a similar 'trimming' process when developing his pcbs, so you might take another look at his method of doing the ground connections - altho some of the other tracking ideas may seem too 'fussy', they leave the options of doing a number of different things with the same pcb.
The type of resistor used for R13 degen next to the jfet will be quite noticeable to the sound quality - there's room enough to add donuts for a 'standard' resistor for the option
I would do the same thing for C3,4 & 5.
On the pcb design pckage, I would increase the donut size a lot to withstand various component changes - possibly increase the track widths also, particularly the high current ones.
There has been some discussion about adding gate stoppers to the input jfets - some people find them useful, others say it's wrong - again, options to find out for yourself ...
s there's room for more options for the power resistors on the Fets, suggest adding extra donuts here also - there are many 'standard' power resistors available and also, 2 in parallel work quite well - there's plenty of room ...
I would suggest another look at the grnd tracking system - perhaps Andrew could add some words here about this and possible grnd plane options
I would separate the Gnd point for the 68uF supply caps from the signal gnd point but not sure where you want to place the amp's central gnd point for chassis earth (or thermistor float?) switcher supplies, speaker returns, signal gnd, etc
Curiously, there are many pcbs with amp cct gnd, signal 0v, speaker returns, etc connected at one point and it works just fine, so this 'separate supply bypass gnd' is just another option.
Oh, just a small thing - suggest adding an extra donut for the P1, 2 as some come with pins 'in line', not 'offset' - an easy option.
If you wanted to, you could reverse the connections to R1 and the input terminal could be arranged for a socket or terminal block with the input gnd ...
I hope some of this could be of some use to you - just suggestions, mind you ....
James,
I'm making a few modifications based on your very helpful observations. Since this board is mostly just for me, I'm not going nuts with all kinds of options, since I know what parts I have on order. But adding some additional donuts takes two seconds and always makes life easier. I'll roll the dice with the fan regulator... if it causes any problems, I can easily disconnect it, and run it from another supply. A few extra donuts will make that adjustment easy. Also, I widened some of the traces as you recommended. After all, I have a lot of room. I upped the main rails and outputs to 5.08mm, which should be pretty good, I think.
You mention grounding scheme, and that is the most interesting thing to me at the moment. I've been doing some reading here on DIY to see some recommendations. I work in industrial automation, so I have familiarity with safety grounding in general, but ground loops and noise is special consideration of amplifiers of course. Attached is how I had planned to ground everything. I've got a separate PSU chassis housing a switcher with a speakOn connecting to the amp chassis. Make sense?
Not sure I understand your comment about separating the gnd point of the 68uF supply caps. Where else would I return them to besides audio gnd? ...Earth? Please clarify if possible.
Thanks
I'm making a few modifications based on your very helpful observations. Since this board is mostly just for me, I'm not going nuts with all kinds of options, since I know what parts I have on order. But adding some additional donuts takes two seconds and always makes life easier. I'll roll the dice with the fan regulator... if it causes any problems, I can easily disconnect it, and run it from another supply. A few extra donuts will make that adjustment easy.
Also, I widened some of the traces as you recommended. After all, I have a lot of room. I upped the main rails and outputs to 5.08mm, which should be pretty good, I think.You mention grounding scheme, and that is the most interesting thing to me at the moment. I've been doing some reading here on DIY to see some recommendations. I work in industrial automation, so I have familiarity with safety grounding in general, but ground loops and noise is special consideration of amplifiers of course. Attached is how I had planned to ground everything. I've got a separate PSU chassis housing a switcher with a speakOn connecting to the amp chassis. Make sense?
Not sure I understand your comment about separating the gnd point of the 68uF supply caps. Where else would I return them to besides audio gnd? ...Earth? Please clarify if possible.
Thanks
Last edited:
I prefer that you make and attach jpeg instead of PDF files.
I can view a jpeg without downloading it, a PDF goes into my download file, if you can't, no problem. You might not be aware of the difference in viewing your attachments, so I thought I would mention it.
Rush
Ah, grounding schemes, eh!
Well, from your diagram, it appears that you want to add the amp case/chassis to the single gnd wire connected to the power supply, yes?
In fact, all the gnd connections in the amp are to go thru this single Gnd wire and that's where the problems occur in that you have the large power supply currents (both channels) and the speaker returns, plus the small signal currents all trying to get back to the 'Zero' gnd point on the external supply via 1 wire and this is bound to produce 'sound degradation', IMO.
I suggest, for a start anyway, that you separate the amp's chassis/heatsinks from the amp's central Zero volt point and use an extra wire for this.
Then I'd think about using a separate gnd wire for each channel back to the cenral point on the supply.
Then I'd run an extra wire for the speaker return (again, each channel) altho plenty of amps don't do this without problems
Then I'd add another separate wire for the gnd connection for those bypass caps (the 68uF ones) on the amp board, hence the change to the existing pcb to allow this.
So, you have each channel amp's central Ov, speaker return 0v, 68uF 0v, plus a separate chassis/heatsink connection to the supply central gnd point, not the CL60 floating earth.
This is one of the problems of using a separate supply case - I would consider sitting the supply in the same case actually - some of the top quality amps are doing this and no problems with 'switcher supply' if screening done well.
I don't know much about switching supplies but I was reading on one of the threads about adding an extra large supply cap to the output of the supply rails - mnot sure if this applies with your particular system.
I haven't seen a choke on the amps supply rails with the bypass caps, so this is something different.
Those filter cap values on the amp brd are a bit large, IMO, but better to have the room and easily change this if you even need them later on.
I hope my 'meanderings' are of some use to you -
My amps generally end up with extra bits squeezed between components, underneath, etc and track cutting, mods and things are quite routine here!
But it's quite surprising how a bit of thought, extra effort, etc can make quite a significant difference to a simple amp diy build
All the best.
Well, from your diagram, it appears that you want to add the amp case/chassis to the single gnd wire connected to the power supply, yes?
In fact, all the gnd connections in the amp are to go thru this single Gnd wire and that's where the problems occur in that you have the large power supply currents (both channels) and the speaker returns, plus the small signal currents all trying to get back to the 'Zero' gnd point on the external supply via 1 wire and this is bound to produce 'sound degradation', IMO.
I suggest, for a start anyway, that you separate the amp's chassis/heatsinks from the amp's central Zero volt point and use an extra wire for this.
Then I'd think about using a separate gnd wire for each channel back to the cenral point on the supply.
Then I'd run an extra wire for the speaker return (again, each channel) altho plenty of amps don't do this without problems
Then I'd add another separate wire for the gnd connection for those bypass caps (the 68uF ones) on the amp board, hence the change to the existing pcb to allow this.
So, you have each channel amp's central Ov, speaker return 0v, 68uF 0v, plus a separate chassis/heatsink connection to the supply central gnd point, not the CL60 floating earth.
This is one of the problems of using a separate supply case - I would consider sitting the supply in the same case actually - some of the top quality amps are doing this and no problems with 'switcher supply' if screening done well.
I don't know much about switching supplies but I was reading on one of the threads about adding an extra large supply cap to the output of the supply rails - mnot sure if this applies with your particular system.
I haven't seen a choke on the amps supply rails with the bypass caps, so this is something different.
Those filter cap values on the amp brd are a bit large, IMO, but better to have the room and easily change this if you even need them later on.
I hope my 'meanderings' are of some use to you -
My amps generally end up with extra bits squeezed between components, underneath, etc and track cutting, mods and things are quite routine here!
But it's quite surprising how a bit of thought, extra effort, etc can make quite a significant difference to a simple amp diy build
All the best.
James- I definitely catch your drift, with the different ground wires. Basically, you are grouping by function each '0V point,' and having a dedicated wire for each. Then instead of having the "star point" inside the amp chassis, you are proposing running them all back to the PSU itself. Although this sounds like a major pain, it could be done rather easily with an 8 conductor speakOn connector.
I understand the recommendation to build everything in one case, and save some hassle. At the same time, there may be some advantages to two cases. I'll think on that one.
At the very least, you've directed my attention to the grounding, which I think I was taking for granted. Since I am designing my own boards, I can add some flexibility in there, so that if I have to make some changes, it doesn't turn into a butchered mess.
BTW, as for the inductor on the rail- that was taken from the application note on page 29 from the manufacturer's design guide. We'll see how it goes.
http://www.vicorpower.com/webdav/si...tions_manual/DesignGuideAppsManual_200J00.pdf
I understand the recommendation to build everything in one case, and save some hassle. At the same time, there may be some advantages to two cases. I'll think on that one.
At the very least, you've directed my attention to the grounding, which I think I was taking for granted. Since I am designing my own boards, I can add some flexibility in there, so that if I have to make some changes, it doesn't turn into a butchered mess.
BTW, as for the inductor on the rail- that was taken from the application note on page 29 from the manufacturer's design guide. We'll see how it goes.
http://www.vicorpower.com/webdav/si...tions_manual/DesignGuideAppsManual_200J00.pdf
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.