Just read the data sheet.
It gives you all the information you need to do this correctly, including which cap to use :
(Brand, size, series, value, .....)
And even a proper PCB layout.
There is no short cut.
And none of the "commercial" implementation for DIY I know of follows those instructions.
(e.g. Murata GRM is by no means the same as Murata GJ8)
Probably to save cost .....
Search this forum under user "Gerhard" and keyword "LT3042".
Tons of useful information, including many comparison messurements.
In the end we make our own to do it properly.
Cheers,
Patrick
Terry,
Yes, I do agree, about the reference filter capacitors.
And this is why I had left space on the board. The pad size is 1206 for both the reference resistor and the filter caps. I had just bought 1uF /1206 PMLCAP types. (have not seen them before).
Presently there are 4,7uF / or 10uF mlcc X7R (TDK) in that position.
The output filter cap is 1812 size, those are 10uF /16V Rubycon PMLCAPs in the photo.
So there is a possibility to assemble 'economical', X7R versions, or full acrylic versions.
Also, the back serves for cooling, a heatsink can be glued on.
The board is 4 layer, and there is a lot of metal in the mass planes.
Still, the size is not big so one should be careful about the thermal load.
Ciao, George
Yes, I do agree, about the reference filter capacitors.
And this is why I had left space on the board. The pad size is 1206 for both the reference resistor and the filter caps. I had just bought 1uF /1206 PMLCAP types. (have not seen them before).
Presently there are 4,7uF / or 10uF mlcc X7R (TDK) in that position.
The output filter cap is 1812 size, those are 10uF /16V Rubycon PMLCAPs in the photo.
So there is a possibility to assemble 'economical', X7R versions, or full acrylic versions.
Also, the back serves for cooling, a heatsink can be glued on.
The board is 4 layer, and there is a lot of metal in the mass planes.
Still, the size is not big so one should be careful about the thermal load.
Ciao, George
Last edited:
Good question - I think there is at least a small effect, but I've barely seen it mentioned or measured. I think the number of layers and dielectric layer thickness affects this. I saw a PDF years ago with some data from Kemet I think that seemed to indicate the effect was worse when pushing the limits - so the smallest package sizes for a given value.
The larger sizes also hold up better in terms of capacitance vs DC bias.
It may be worth investigating if those with "soft" boardflex terminations help, or the caps with the metal L shaped terminations.
How to reduce acoustic noise of MLCCs in power applications - Power management - Technical articles - TI E2E support forums
Just read the data sheet.
It gives you all the information you need to do this correctly, including which cap to use :
(Brand, size, series, value, .....)
And even a proper PCB layout.
There is no short cut.
And none of the "commercial" implementation for DIY I know of follows those instructions.
(e.g. Murata GRM is by no means the same as Murata GJ8)
Probably to save cost .....
Search this forum under user "Gerhard" and keyword "LT3042".
Tons of useful information, including many comparison messurements.
In the end we make our own to do it properly.
Cheers,
Patrick
I might add that the LT evaluation manual for these regulators has good information that is not in the datasheet.
If you want current from the LT3042 family you don't get far with cooling attempts.
Use it with an external pass transistor. The circuit is in the data sheet.
Noise performance is about the same as without the pass transistor.
< Lt3042_DH44 | Low noise regulator consisting of LT3042 and D… | Flickr >
Noise is on the next picture to the left. The steep rise at low frequencies goes on the
preamplifier.
Joseph K, if you have that thing with the 20 ADA4898, the input capacitor
must be MUCH larger. 2200..4700u is about right. That will be probably no foil cap.
You may get away with Al electrolytics, I ended up with wet slug tantalum
for $$$. Organic polymer is too leaky. The change removes the steep noise
rise of the amplifier. Short the amplifier side of the cap when connecting to
a large DC input voltage. The op amps will thank you for it. Remove the short
after connection. Also, do not short the input as long as the cap is charged.
There is still a lot of energy in it that may zener the op amps.
regards, Gerhard
Use it with an external pass transistor. The circuit is in the data sheet.
Noise performance is about the same as without the pass transistor.
< Lt3042_DH44 | Low noise regulator consisting of LT3042 and D… | Flickr >
Noise is on the next picture to the left. The steep rise at low frequencies goes on the
preamplifier.
Joseph K, if you have that thing with the 20 ADA4898, the input capacitor
must be MUCH larger. 2200..4700u is about right. That will be probably no foil cap.
You may get away with Al electrolytics, I ended up with wet slug tantalum
for $$$. Organic polymer is too leaky. The change removes the steep noise
rise of the amplifier. Short the amplifier side of the cap when connecting to
a large DC input voltage. The op amps will thank you for it. Remove the short
after connection. Also, do not short the input as long as the cap is charged.
There is still a lot of energy in it that may zener the op amps.
regards, Gerhard
Some measurements appeared in the thread on this supply: https://www.diyaudio.com/forums/pow...gulator-based-studer-900-a-7.html#post5716864
Hi, jonners, which post # are you referring to? Do you know the supplier of the board being measured?
Hi. I'm referring to post #68. There are some more measurements in post #103.
As to the supplier of the boards, you would have to ask those who posted.
However the boards are readily found on eBay by searching for 'Studer900 power supply'.
Edit: Rather than continue here off-topic I suggest going to the other thread for any further discussion of the Studer clone.
Last edited:
I have 2 from this seller powering dac IV stages. Seem very good for the cost. I haven't taken measurements.
￡12.59 32%OFF | Lusya STUDER900 Regulator Power supply board 5-28V Can assembled into double power board DIY kit A3-006
AliExpress
￡12.59 32%OFF | Lusya STUDER900 Regulator Power supply board 5-28V Can assembled into double power board DIY kit A3-006
AliExpress
I got this one from e-bay which shows the same series of pictures. The manufacturer is listed as SODIAL.
5-28V STUDER900 Regulator Power Supply Board Can Assembled Into Double Power 4F3 194452680023 | eBay
I have this on hand waiting to be assembled, but not the LT3045/LT3094 PSU yet.
So which one of those EBay low noise units I should buy? The requirements are 18V and 1A (minimum I guess) The PSU wold power Antelope Zodiak Gold DAC and replace wallmart switching gizmo. While DIY has it charms, $100 or so for a nice looking box and a lot of time saved (dedicated to making $$) have something to go for as well. Antelope charges $1200 for their Volticus PSU which is kind of ridiculous considering the unit is manufactured in cheap labor country ..
For the cost of them you could try both?
The Studer 900 does come in a single supply version which is all you need and is cheaper. Unless you want to build it yourself I was also advise the completed psu....they take quite some time to populate and solder and the cost difference is negligible.
You could always upgrade from one of these to a Salas design such as the Ubib should you see the need to try something potentially better.
The Studer 900 does come in a single supply version which is all you need and is cheaper. Unless you want to build it yourself I was also advise the completed psu....they take quite some time to populate and solder and the cost difference is negligible.
You could always upgrade from one of these to a Salas design such as the Ubib should you see the need to try something potentially better.
Last edited:
Do not buy unless you need a +/-5V dual output. The item description specifies a range of output voltage up to +/-15V making it sound like user adjustable. I expect to receive several resistors to pick the right value for the output voltage that I want. I got a fixed +/-5V board which is not adjustable. Yes, it is a dual output 5V board and completely useless to me. I need +/-15V for a phono op-amp board.
Last edited:
I am writing to the LT3045/LT3094 PSU seller to try to get one with +/-15VDC output. I received one with +/-5VDC.
I see a resistor with 'adj' silkscreened next to it. I assume this is the voltage setting R. For lt3045 I believe that value should be 150k for 15v, according to the data sheet.
Maybe the negative rail will be the same.
**edit...I also now see they list a few values of R fir common output voltages. I guess you just got an incorrectly set up psu board.
Are those adj resistors tiny tiny?
Maybe the negative rail will be the same.
**edit...I also now see they list a few values of R fir common output voltages. I guess you just got an incorrectly set up psu board.
Are those adj resistors tiny tiny?
Last edited:
The "adj" resistor are tiny surface mount one for both rail. The seller advertised a output range which I expect to select by soldering a jumper bridge. Or pick a resistor according to the output I need to solder it on. The board as received is a fixed output +/- 5V, not +/- 15V as I requested.
The seller offered a $10 "compensation". I did not accept because the +/- 5V PSU is of no use to me.
I will focus on finish building the Studer 900 PSU and see how well it works. It has received good review and showed good measurement from DIYAudio members. I got a kit. I will measure the transistors before I solder them on. Most substitute, if needed, are readily available from my bench except the output MJE15034 which is in stock at Mouser and Digi-Key.
Last edited: