Bipolar SMPS amperage rating question

[Potato Cakes]

Hello,

I have a single power source audio circuit that am incorporating with a couple of other circuits using ICs, so I am using a bipolar SMPS to power the whole unit, running the single power source (+V, 0V) circuit off of the +V,-V. My initial testing with this scheme works great, but issue I am running into is the SMPS is sensing the current draw of the whole unit is too much, causing it to cycle on/off. The power supply is rated for 700mA for V+ and 700mA for V-. The current draw for the entire unit is about 360mA, so I should be well with in the capacity of the power supply. When I rewire the single source circuit to be +V and 0V, the power supply works fine. I tested with a bigger SMPS that was double the current rating and it worked fine with the bipolar scheme mentioned above.

My question pertains solely to the SMPS amp rating. My understanding is that the 700mA listed on the power supply pertains to +V to 0V and -V to 0V. Would the amp rating be different between +V and -V? That seems to be the case with what I am experiencing but I do not quite understand why.

Thanks!

Paul

 

[ruffrecords]

Exactly which SMPS are you using?

Cheers

Iam

 

[abbey road d enfer]

Hello,

I have a single power source audio circuit that am incorporating with a couple of other circuits using ICs, so I am using a bipolar SMPS to power the whole unit, running the single power source (+V, 0V) circuit off of the +V,-V. My initial testing with this scheme works great, but issue I am running into is the SMPS is sensing the current draw of the whole unit is too much, causing it to cycle on/off. The power supply is rated for 700mA for V+ and 700mA for V-. The current draw for the entire unit is about 360mA, so I should be well with in the capacity of the power supply. When I rewire the single source circuit to be +V and 0V, the power supply works fine. I tested with a bigger SMPS that was double the current rating and it worked fine with the bipolar scheme mentioned above.

My question pertains solely to the SMPS amp rating. My understanding is that the 700mA listed on the power supply pertains to +V to 0V and -V to 0V. Would the amp rating be different between +V and -V? That seems to be the case with what I am experiencing but I do not quite understand why.

Thanks!

Paul

What nominal voltage for teh smps. What are you powering? In particular what piece are you trying to power between V+ and V-?

 

[Potato Cakes]

Exactly which SMPS are you using?

Cheers

Iam

It is branded as Heritage Audio and the model number is PSU-02, which is used to power 500 series lunch boxes. The actual manufacturer is a Taiwanese company that does OEM. It is rated for 700mA for +17.5V, 700mA for -17.5V and 80mA for 48V.

Thanks!

Paul

 

[Potato Cakes]

What nominal voltage for teh smps. What are you powering? In particular what piece are you trying to power between V+ and V-?

The unit I am building consists of a 1073 with two additional BA283AV circuits, a headphone amp that uses NE5532's and an unbalancing circuit that uses INA134's. I had this setup with a PSU that was +24V and 0V, but to accommodate the IC's I had to do all these passive rail splitters plus I never found a good way to convert +24V to 48V for phantom power. I decided I needed a simpler way to do the power section that's when I ended up with a bipolar PSU solution. SMPS units that are used for 500 series units have become very good and even more so with some simple filtering. So I did some experimenting with just a BA283AV circuit with one of these PSU's and I found that even when using it as a preamp with dynamic microphone, I got a very clean and clear signal, completely free of the typical noise that can happen with this circuit, even with the PCB outside of a case. So I am proceeding with this bipolar design and using a regulator to drop the -V rail about 6V per Winston O'Boogie's recommendation on a different thread.

According to my calculations based off of IC spec sheets and what I am measuring on the BA cards, a 700mA PSU should be plenty for this build, but I don't know why 700mA between V+/0V and V-/0V would be different between V+/V-. I tried numerous wiring configurations and this where I ended up last night, which prompted me to check in here.

Thanks,

Paul

 

[JohnRoberts]

Have you measured your actual current draw?

JR

 

[Bo Deadly]

That cycling on / off is protection mode.

Do you have a big capacitor on the output of the SMPS? If the cap is too big, it will draw too much current on startup and cause the output to go into protection mode. SMPS outputs have very little noise compared to rectified AC so you don't need huge filter caps like in a linear supply. Just reduce the size of the cap until it starts up ok.

Alternatively, the proper way to filter an SMPS (IMO) is to use a capacitance multiplier (CM) [1]. The CM acts like a larger cap but it's actually not in two ways: one is that it's actually not buffering current so it's not like you miraculously get a big cap from a little one but second it is slow to "charge up" in that the RC of the CM is slow compared to a regular cap. So that slow turn on also keeps the SMPS from going into protection mode. Then you can follow the CM with a large cap to really make the source Z low at all frequencies.

For smaller loads like a couple hundred mA of phantom power, a simple LC filter might be sufficient to just get the high frequency noise. But inductors are big and expensive so my feelings about this have recently evolved and I'm starting to think that a simple CM is the way to go. Although for high currents you need the two transistor version.

 

[Potato Cakes]

Have you measured your actual current draw?

JR

I get about 90mA per BA283AV (approximately 2V across a 22ohm resistor). I haven't been able to get a reading over the entire circuit because I was running into the problem described above, so I measured the circuits individually. I do know that I get about 150mA for just the 1073 circuit and when I add the additional BA283AV cards that would math wise put me around 330mA, which fits within 360mA proximation with the additional IC chips. I also should add that this is a fully functional and working build. The reason for switching to a bipolar design is have a better way to accommodate the IC's and phantom power.

Thanks!

Paul

 

[Potato Cakes]

That cycling on / off is protection mode.

Do you have a big capacitor on the output of the SMPS? If the cap is too big, it will draw too much current on startup and cause the output to go into protection mode. SMPS outputs have very little noise compared to rectified AC so you don't need huge filter caps like in a linear supply. Just reduce the size of the cap until it starts up ok.

Alternatively, the proper way to filter an SMPS (IMO) is to use a capacitance multiplier (CM) [1]. The CM acts like a larger cap but it's actually not in two ways: one is that it's actually not buffering current so it's not like you miraculously get a big cap from a little one but second it is slow to "charge up" in that the RC of the CM is slow compared to a regular cap. So that slow turn on also keeps the SMPS from going into protection mode. Then you can follow the CM with a large cap to really make the source Z low at all frequencies.

For smaller loads like a couple hundred mA of phantom power, a simple LC filter might be sufficient to just get the high frequency noise. But inductors are big and expensive so my feelings about this have recently evolved and I'm starting to think that a simple CM is the way to go. Although for high currents you need the two transistor version.

I did think it might be the power reservoir capacitors on the cards as I had a multichannel preamp build that had the same issue and I had to use a smaller values for those caps from what the schematic showed. So I pulled those caps just to check and I had the same issue.

I have not incorporated any filtering at this point as I am trying to get over the PSU issue first. I do have the larger SMPS that I think will fit in the build which would solve the issue, but my question still remains the same in regards to ratings for SMPS. From what I am measuring and reading from spec sheets, 700mA should be more than enough to sufficiently power this build, which for some reason doesn't when using the V+/V-.

Thanks!

Paul

 

[JohnRoberts]

Ok next test is to load the suspect PS with passive resistor loads (may require power resistors), to see if the connection manner matters.

Note: if one PS is sourcing current that the other supply is sinking, their grounds need to be bonded together so currents can find their way home.

JR 

 

[Potato Cakes]

Ok next test is to load the suspect PS with passive resistor loads (may require power resistors), to see if the connection manner matters.

Note: if one PS is sourcing current that the other supply is sinking, their grounds need to be bonded together so currents can find their way home.

JR

I start with this process and it didn't like it when I approached 100 ohm, which I believed to be the start of messing with the reference for V+. As I mentioned, for the single source audio circuit, the V- is being used as the 0V, which makes the difference between V+ and V- a "better" voltage rail more suitable to that particular circuit than just the V+ and 0V that is coming off of the PSU. There is no electrical ground connection per se.

As for bonding the grounds, both the V+ and V- are referencing the same 0V and chassis ground per the PSU labels. I don't have the schematic and it is a sealed design, so I'm going off the manufacturer's labeling.

Thanks!

Paul

 

[ruffrecords]

I think this may be due to using a single supply circuit between +V and -V rails. The problem could be caused by the fact that the ground as far as the single supply circuit is concerned is -V but as far as the other circuits connected to the bipolar supply are concerned it is the common 0V of the bipolar supply. You may therefore be introducing a low impedance path between -V and the supply 0V which is causing the overload. The SMPS is probably fine.

Cheers

Ian

 

[Potato Cakes]

I think this may be due to using a single supply circuit between +V and -V rails. The problem could be caused by the fact that the ground as far as the single supply circuit is concerned is -V but as far as the other circuits connected to the bipolar supply are concerned it is the common 0V of the bipolar supply. You may therefore be introducing a low impedance path between -V and the supply 0V which is causing the overload. The SMPS is probably fine.

Cheers

Ian

One of the things I've yet to implement is a regulator to drop the V- going to the single source and hopefully it will act as a buffer to the IC's that are bipolar. I'll report back.

Thanks!

Paul

 

[Matador]

I think this may be due to using a single supply circuit between +V and -V rails.

Agreed.  I bet he can replicate this with a test load across V+ and V-, versus two 700mA test loads, one for each rail.

It's asking the SMPS's negative rail to sink current, which might confuse the regulation circuitry to believe that the circuit is in a current limit situation.