Guys, I am in the process of modding a squeezebox to blend seamlessly with my tda1541a dac. One of the remaining subjects is the power supplier to the squeezebox.
I am not using the squeezebox dac, nor am I interested in the WiFi functionality. I need a 250mA@3.3V and a 350?mA@1.2V supply for the remaining (digital) parts.... any suggestions for low noise low impedance over a wide bandwidth, yet able to supply the requested currents and not too bulky/hot ?
I am not using the squeezebox dac, nor am I interested in the WiFi functionality. I need a 250mA@3.3V and a 350?mA@1.2V supply for the remaining (digital) parts.... any suggestions for low noise low impedance over a wide bandwidth, yet able to supply the requested currents and not too bulky/hot ?
To buy? To make? What complexity level or budget?
What makes you feel it needs wide bandwidth? For that matter if not using the DAC, what makes you feel digital circuits need low noise?
Is it built into the squeezebox or external? Using squeezebox power source or a separate transformer?
Offhand the simplest solution would seem to be providing 5VDC input to a low dropout linear regulator. There will be heat but a manageable level at 0.4W /3.3V, 1.3W / 1.2V. 1.3W is high enough there should be a modest heatsink on the regulator providing 1.2V. One example of a low dropout regulator is LT1084, though there are probably newer, higher performance regulators today, with more care needed in circuit design to achieve the improved performance.
What makes you feel it needs wide bandwidth? For that matter if not using the DAC, what makes you feel digital circuits need low noise?
Is it built into the squeezebox or external? Using squeezebox power source or a separate transformer?
Offhand the simplest solution would seem to be providing 5VDC input to a low dropout linear regulator. There will be heat but a manageable level at 0.4W /3.3V, 1.3W / 1.2V. 1.3W is high enough there should be a modest heatsink on the regulator providing 1.2V. One example of a low dropout regulator is LT1084, though there are probably newer, higher performance regulators today, with more care needed in circuit design to achieve the improved performance.
I'm using my own fet based regs in mine for the 3.3v and the 5v. Problem is they need quite a lot of voltage headroom. This is ok if you feed multiple psu lines into the touch. You could try the new fidelity audio SPower 'D' which I believe only needs about 700mV. I would still look to utilise an additional supply line in because if the reg was originally supplied via the internal smps, it's really fussy. I tried to run a 3.3v reg from the 9v smps rail but it shuts to smps rail down to about 5v. It does the same if you add too much capacitance across the internal regs which are supplied by this smps rail. If the reg is supplied directly from the main 5v supply, you'll be ok.
This is the test 5v reg I have on the DAC. I had to change the FET I use to smt due to space so it's a little less tidy than I'd hoped.
This is the test 5v reg I have on the DAC. I had to change the FET I use to smt due to space so it's a little less tidy than I'd hoped.
An externally hosted image should be here but it was not working when we last tested it.
You can also rig a linear regulator using a TLV431 and an external transistor (1.24V reference) to generate the voltages from the 5V rail. See the TI application note. If you still have a question I am sure someone here can help.
The LED's (AndrewT's suggestion) will work. The TLV431 will just have an advantage on temperature and Vref stability.
Tony
The LED's (AndrewT's suggestion) will work. The TLV431 will just have an advantage on temperature and Vref stability.
Tony
Use some small LDO devices, from Ti/National or similar, there are plenty to choose from as adding these voltages to local areas of digital circuitry or for the outputs of an FPGA is very very common these days. Ti/Nat also has a workbench so you can try different topographys to suit your buget, available space, efficiency etc etc
One thing I'll mention if you're trying this... lots of devices that run off multiple rails have specific power sequencing requirements. I wouldn't be surprised if there's a chip on the board that could suffer damage if 1.2V comes up before 3.3V.
I'd take an inventory of chips on the board and make sure this isn't the case. Enforcing this means whatever power supply you come up with needs power sequencing built in.
I'd take an inventory of chips on the board and make sure this isn't the case. Enforcing this means whatever power supply you come up with needs power sequencing built in.
Are you referring specifically to the squeezebox, or a general case where latch-up is a problem?
It have not found (on the web) any references to power up sequencing being a problem on the SB.
What other rails have you identified?
I've done a fair bit but not come across this one so far. The PSU is really quite strange. I dont really understand why the main rail is only 5v.
The internal DAC runs on 5v. The wall wart smps is 5v. Inside the unit goes into another smps up to approx 10v to drive a 5v reg for the dac and touch screen sense.
All looks a bit strange with multiple smps inside!!!
BTW, I've got a total of 5 low noise regs (2 on clocks) and I have no sequencing issues.
I've done a fair bit but not come across this one so far. The PSU is really quite strange. I dont really understand why the main rail is only 5v.
The internal DAC runs on 5v. The wall wart smps is 5v. Inside the unit goes into another smps up to approx 10v to drive a 5v reg for the dac and touch screen sense.
All looks a bit strange with multiple smps inside!!!
BTW, I've got a total of 5 low noise regs (2 on clocks) and I have no sequencing issues.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.