New Regulator
- Thread starter Bemused
- Start date
RyanSoundLab
Trade: Ryan Sound Lab
Tony,
I think there's a wrong connection on......just kidding. With Eagle, it's hard to make a mistake. The Gerbers look very good.
Kit
I think there's a wrong connection on......just kidding. With Eagle, it's hard to make a mistake. The Gerbers look very good.
Kit
RyanSoundLab
Trade: Ryan Sound Lab
Looking good, Tony. It always amazes me how such professional work can be done for so little. I can make boards at home but the solder mask (green stuff) is almost impossible to do properly and critical to a nice job (yes, I've seen the threads on how to do it yourself). Beautiful work. Hope the circuit is now stable and performing to expectations. We await your results.
Kit
Kit
Revision r032 Stuffed and ready for testing.
Thought I had de-fluxed until I saw this picture.
Improved scope measurement has the scope lead welded on the board with a series 100R. About the smallest loop area I can get with the coax wrapped around a 1206 SMD
Well would be ready if I had not brought a BNC socket sized for a welding cable.
Grrrrr
Thought I had de-fluxed until I saw this picture.
Improved scope measurement has the scope lead welded on the board with a series 100R. About the smallest loop area I can get with the coax wrapped around a 1206 SMD
Well would be ready if I had not brought a BNC socket sized for a welding cable.
Grrrrr
RyanSoundLab
Trade: Ryan Sound Lab
Tony,
Very nice soldering work for SMD's using an iron! Really well done. You even took the trouble to line up the resistor printing.
For really clean work, I use "water soluble flux" core solder. It's available in both lead free and regular (which is much easier to work with, of course). When you're done soldering, you run the board under hot water and scrub with a toothbrush (no kidding), then blow dry to get the remaining water off. (I use an air compressor). It comes out perfectly clean. The components seem to have no trouble with the water bath. I use it all the time on my RSL boards. If you use the alcohol-based sprays, it's almost as good but you go through a lot of alcohol.
Agree with the coax size! You can use a regular probe if you solder a short piece of wire to the contact point(s) of interest and then hook the probe onto them. Might be easier and since this is not a RF project, should be quite adequate.
Listening test next?
Kit
Very nice soldering work for SMD's using an iron! Really well done. You even took the trouble to line up the resistor printing.
For really clean work, I use "water soluble flux" core solder. It's available in both lead free and regular (which is much easier to work with, of course). When you're done soldering, you run the board under hot water and scrub with a toothbrush (no kidding), then blow dry to get the remaining water off. (I use an air compressor). It comes out perfectly clean. The components seem to have no trouble with the water bath. I use it all the time on my RSL boards. If you use the alcohol-based sprays, it's almost as good but you go through a lot of alcohol.
Agree with the coax size! You can use a regular probe if you solder a short piece of wire to the contact point(s) of interest and then hook the probe onto them. Might be easier and since this is not a RF project, should be quite adequate.
Listening test next?
Kit
Thanks for the flux tips Kit, and the complement.
Learned that aligning components pays dividends when the thing won't work, especially when dev revs are covered by components. Spent a while looking at / copying our Mister6 s work on Starfish years ago, man can he just dress a through hole resistor.
Not powered the board up yet!
Some new footprints may be wrong amongst the usual problems, so bench test, checking DC operating points and general behaviour befor I toast any more Sabres. I've learnt to slow down and resist my urge to get things in circuit.
The lack of a suitable BNC connecter has given me an excuse to slow down.
Was taking PDs advice on improving scope work from earlier in thread.
Been thinking of giving your frying pan job a whirl but I don't enjoy hand applying paste so may try a twist.
Sort of hybrid of what I currently do and the frying pan. Thinking of stuffing with the iron as usual but going easier on the solder, then refluxing and then into to oven, sort of expexting to get better auto alignment of the parts.
Learned that aligning components pays dividends when the thing won't work, especially when dev revs are covered by components. Spent a while looking at / copying our Mister6 s work on Starfish years ago, man can he just dress a through hole resistor.
Not powered the board up yet!
Some new footprints may be wrong amongst the usual problems, so bench test, checking DC operating points and general behaviour befor I toast any more Sabres. I've learnt to slow down and resist my urge to get things in circuit.
The lack of a suitable BNC connecter has given me an excuse to slow down.
Was taking PDs advice on improving scope work from earlier in thread.
Been thinking of giving your frying pan job a whirl but I don't enjoy hand applying paste so may try a twist.
Sort of hybrid of what I currently do and the frying pan. Thinking of stuffing with the iron as usual but going easier on the solder, then refluxing and then into to oven, sort of expexting to get better auto alignment of the parts.
RyanSoundLab
Trade: Ryan Sound Lab
Auto alignment isn't a sure thing. The larger components (tantalums, for example) have to be completely hot and even then it's iffy. I think it depends on how you do it. If I put the boards/parts in a frying pan (usual method), they heat from the bottom and only the pads and solder gets very hot (not the components) so they don't self-align much. If I use a hot air gun and hit the parts from the top down with hot air, they seem to "jump" into place as you would like. This is much more time-consuming and you can even blow the parts off the boards if you're not careful but they look better aligned.
A third method, which is closer to a real factory job, is using a toaster oven. Then, everything gets hot and they will align as best as they can (again, the bigger components are harder to get aligned this way unless it's an IC with a million pins (like a Sabre)). That should work very well.
Ultimately, you and those looking at your thread are the only ones who will know. If it works well, that's the real bottom line.
Good luck with testing. Agree on going slow!
Kit
ps. Where are you getting your Sabre chips? I haven't found much availability unless you're a manufacturer.
A third method, which is closer to a real factory job, is using a toaster oven. Then, everything gets hot and they will align as best as they can (again, the bigger components are harder to get aligned this way unless it's an IC with a million pins (like a Sabre)). That should work very well.
Ultimately, you and those looking at your thread are the only ones who will know. If it works well, that's the real bottom line.
Good luck with testing. Agree on going slow!
Kit
ps. Where are you getting your Sabre chips? I haven't found much availability unless you're a manufacturer.
Bench testing,
All DC operating points seem good.
Scoping output.
Very low level signal on scope around <4mV pk-pk at 10MHz and 100uV at 100MHz.
Influenced by lights, computers, screens etc.
Touching or moving the scope lead generally attenuates.
Moved from conventional scope lead to the soldered in place lead with loop areas of a just few millimetres and suddenly all is rock steady and flat lining.
Now immune to touching except around the CCS which causes a little wide band noise and a momentary scope trace bump.
Feeling happy
5mV and 0.5uS per division.
All DC operating points seem good.
Scoping output.
Very low level signal on scope around <4mV pk-pk at 10MHz and 100uV at 100MHz.
Influenced by lights, computers, screens etc.
Touching or moving the scope lead generally attenuates.
Moved from conventional scope lead to the soldered in place lead with loop areas of a just few millimetres and suddenly all is rock steady and flat lining.
Now immune to touching except around the CCS which causes a little wide band noise and a momentary scope trace bump.
Feeling happy
5mV and 0.5uS per division.
Thanks again.
Probably not worth me messing around any further as I find the iron trouble free and fast but I now have an itch
Yes I strip more than the Sabre when swapping them out with a hot air dryer.
I have satisfactorily used a dremmel to cut sabre legs before removal with an iron before now but still blow the bypass caps away when I refit with the hot air.
I don't buy often, if I feel flush I have a new Buffalo and I have had a failure on one swap out and then its getting both expensive and frustrating with a goofed swap out.
Last dead Sabre was fixed with a new Buffalo.
When I do purchase a Sabre its just one off. I use the UK distributor, telephone first, get passed to the PA of someone whom I forget the name then follow with email and a BACS payment. Previously taken a day or two for delivery. Comes in a posh little black box.
Tony
RyanSoundLab
Trade: Ryan Sound Lab
Interesting. Over here, I have to deal with an ESS sales rep (which I haven't done yet) to get any parts. Not available from any distributor that I know of. One thing I don't need is a sales rep pushing me!
Kit
Kit
Perhaps the difference is that I approached them as a private diy buyer requesting just one sabre. They were around 70UKP delivered so probably at list price or even more?
I know they can be had for 90USD already on a board from HK
I guess they don't use distributers for domestic sales.
I got the distributors name from ESS site.
I know they can be had for 90USD already on a board from HK
I guess they don't use distributers for domestic sales.
I got the distributors name from ESS site.
Regulator is now in the DAC and showing a little on the scope.
Gut feeling is its spurious noise from the DAC as changing the scope cable route accentuates or attenuates the signal.
First photo is with the cable deliberately draped right across a rPI and Arduino.
Second is best exit route I could find.
Lightly gripping the scope lead attenuates further.
Can see some burst firing?
Gut feeling is non of this is regulator problems so still feeling happy.
In circuit start up is good showing 1/100 volt overshoot using a calibrated Fluc as a logger.
5mV and 0.5uS per division.
Gut feeling is its spurious noise from the DAC as changing the scope cable route accentuates or attenuates the signal.
First photo is with the cable deliberately draped right across a rPI and Arduino.
Second is best exit route I could find.
Lightly gripping the scope lead attenuates further.
Can see some burst firing?
Gut feeling is non of this is regulator problems so still feeling happy.
In circuit start up is good showing 1/100 volt overshoot using a calibrated Fluc as a logger.
5mV and 0.5uS per division.
Well still having a little mental trouble signing off on this regulator, a little disappointed that I still see something on the output rail so I furthered the measurement technique by wrapping the scope earth around the scope lead.
Signal is different than yesterday.
Today I cleaned the scope BNC connectors and added the wrapped earth but also found all previous measurements were taken with the scope main earth disconnected, today I connected the main earth.
First shot is with the whole DAC powered down, second shot is with it running.
Remaining pictures show the measurement technique.
ETA
This regulator output is one of the quietest points in the DAC, I have similar to massively more, horrible noise scoping other points, ground planes etc.
Signal is in the region of 1mV (RMS)
5mV and 0.5uS per division.
Signal is different than yesterday.
Today I cleaned the scope BNC connectors and added the wrapped earth but also found all previous measurements were taken with the scope main earth disconnected, today I connected the main earth.
First shot is with the whole DAC powered down, second shot is with it running.
Remaining pictures show the measurement technique.
ETA
This regulator output is one of the quietest points in the DAC, I have similar to massively more, horrible noise scoping other points, ground planes etc.
Signal is in the region of 1mV (RMS)
5mV and 0.5uS per division.
RyanSoundLab
Trade: Ryan Sound Lab
Tony,
Really hard to conclude your "noise" problem is with your new power supply. Now that I see a picture of the whole setup, there are other possibilities for noise pickup, such as (1) the very long coax scope input cable or (2) running the cable near the board that looks like it's got the DAC on it (with its clock - although you still had the noise pickup with the DAC off). The spread out nature of the whole setup makes RFI from outside sources more likely, too. The only thing you might try to be sure your supply isn't at fault is unsolder the scope connection from the board and just short it to its own ground with a similar value resistor to what's on the board. Position the coax just like it is now. If you still get noise pickup, then you know it's the scope setup and not the board.
Good luck!
Kit
Really hard to conclude your "noise" problem is with your new power supply. Now that I see a picture of the whole setup, there are other possibilities for noise pickup, such as (1) the very long coax scope input cable or (2) running the cable near the board that looks like it's got the DAC on it (with its clock - although you still had the noise pickup with the DAC off). The spread out nature of the whole setup makes RFI from outside sources more likely, too. The only thing you might try to be sure your supply isn't at fault is unsolder the scope connection from the board and just short it to its own ground with a similar value resistor to what's on the board. Position the coax just like it is now. If you still get noise pickup, then you know it's the scope setup and not the board.
Good luck!
Kit