LM723 PSU with 0V lowest voltage

bountyhunter said:

Avoid using Y5F or Z5U dielectric they are complete junk. X5R and X7R are best buy and good quality.

Click to expand...

I tend to avoid X5R too. Why bother with X5R when you can use X7R.

spec

spec said:

I tend to avoid X5R too. Why bother with X5R when you can use X7R.

spec

Click to expand...

some of the large ones were only available in X5R.

bountyhunter said:

some of the large ones were only available in X5R.

Click to expand...

Yes that is true, but when you do a worst case tolerance with large X5R ceramic capacitors, you find that they are not necessarily that large.

spec

I meant the large value ranges.

bountyhunter said:

I meant the large value ranges.

Click to expand...

So did I.

spec

My supplier sells "ceramic", "mks", "mkt", mkp" non-polarised capacitors: which should I (not) choose?

earckens said:

My supplier sells "ceramic", "mks", "mkt", mkp" non-polarised capacitors: which should I (not) choose?

Click to expand...

Capacitors can have a big impact on equipment performance.

My (biased) advice is not to use ceramic capacitors, apart from decoupling/supression. All other non-polarised capacitor types will be OK for frequency compensation on your power supply.

MKS, MKT: polyester metal film capacitors are general purpose types. They have a high temperature rating and are much cheaper than other film or foil capacitors.

PPS: polyphenylene sulfide metalized film capacitors are pretty close to MKP in electrical performance. They are higher temperature (same as MKS/MKT) and lower cost than MKP (used for HiFi applications)

MKP: polypropylene metalized film capacitors are excellent electrically but are lower temperature than MKS/MKT/PPS and not suitable for surface mount and are expensive (used for high current, high voltage applications and audiophile applications)

MFP, polypropylene foil capacitors are superb electrically but are lower temperature (same as MKP) and not suitable for surface mount. They are rare and very expensive (used for very high current, high voltage applications and high end audiophile applications)

Other capacitors: polystyrene, silver mica for example have good low current, high frequency characteristics.

spec

spec said:

Capacitors can have a big impact on equipment performance.

My (biased) advice is not to use ceramic capacitors, apart from decoupling. All other non-polarised capacitor types will be OK for frequency compensation on your power supply.

MKS, MKT, polyester metal film capacitors are general purpose capacitors. They have a high temperature rating and are much cheaper than other film or foil capacitors.

PPS, polyphenylene sulphide metalized film capacitors are pretty close to MKP in electrical performance. They are higher temperature and lower cost (used for HiFi applications)

MKP, polypropylene metalized film capacitors are excellent electrically but are lower temperature than MKS and not suitable for surface mount and are expensive (used for audiophile applications)

MFP, polypropylene foil capacitors are superb electrically but are lower temperature (like MKP) and not suitable for surface mount. They are rare and very expensive (used for high end audiophile applications)

Other capacitors: polystyrene, silver mica for example have good low current, high frequency characteristics.

spec

Click to expand...

You are tha men! I searched the net this morning before posting my question but no proper answer. You and bountyhunter did it. Thank you.

earckens said:

You are tha men! I searched the net this morning before posting my question but no proper answer. You and bountyhunter did it. Thank you.

Click to expand...

No probs earckens,

Non-polarized capacitors types are a nightmare to untangle and new types keep being introduced.

And polarized capacitors are another minefield, with tantalum capacitors, like ceramic capacitors, being the bad/good boys.

spec

https://en.wikipedia.org/wiki/Film_capacitor

Hi, now about current measurements: I need to measure the outgoing current and am considering an INA219B. Can I place its shunt resistor in the negative supply line yet connect it to a controller supplied with the positive fixed voltage rail from the PSU, and the negative from the common negative rail?

Looking at the spec sheet it seems ok but I want to make sure from this forum. i will not use the voltage measurement feature of the chip.

Thank you.

Hi earckens,

I take it that you are aiming at highly accurate current measurement, which means that you would not want to use the voltage drop across R3 in the schematic of post #22.

If you put a current monitoring resistor in the 0V line you will degrade the voltage regulation of the Power Supply Unit (PSU).

So the current monitoring resistor would need to fit at the high side, between R17 and R12. This means that the input to the current monitoring amplifier will experience the full voltage range of the PSU: 0V to 30V. As the INA219 only has an input range up to 26V it would not be easy to use without some fancy circuitry.

This would still leave the inaccuracy caused by the current drain of the voltage sense resistors R12 and R13 though. This could probably be reduced by increasing the values of R12 and R13.

If you just want a voltage input representing current into your controller, rather than SPI (correction I2C), you could use an instrumentation amplifier: Texas Instruments and Analog Devices have some suitable chips. Or perhaps the INA193 series of current monitors which have a wider common mode voltage range of -16V to 80V: https://www.ti.com/lit/ds/symlink/ina196.pdf

spec

Hi Spec, reading your text and subsequently again studying your post 22 (and 12) I discovered an anomaly between what I think the layout should be and what you drafted in your posts. Now I may be wrong, and that is why I would like to ask you if I may pm you the circuit that you had drafted in post 12 with what I think should be the correct layout.
I would appreciate tremendously you taking a look at my drawing and comment on the corrections that I think should be in place. They all are centered around the layout for +IN and -IN.

For the current measurement what do you mean by "SPI"? Installing measurement on the high side makes it a problem to power the controller (Atmega 328). Unless I use a reference zener and hang this from the positive, with a drain resistor to negative and hang the controller on the positive side: POS - CATH Zener - ANODE Zener (connected to negative of the controller) - DRAIN Resistor. The Zener current plus the controller current pass through the drain resistor to GND. Workable?

earckens said:

Hi Spec, reading your text and subsequently again studying your post 22 (and 12) I discovered an anomaly between what I think the layout should be and what you drafted in your posts. Now I may be wrong, and that is why I would like to ask you if I may pm you the circuit that you had drafted in post 12 with what I think should be the correct layout.
I would appreciate tremendously you taking a look at my drawing and comment on the corrections that I think should be in place. They all are centered around the layout for +IN and -IN.

Click to expand...

Sure: I will have a look.

earckens said:

For the current measurement what do you mean by "SPI"?

Click to expand...

My error: I should have said I2C interface.

earckens said:

Installing measurement on the high side makes it a problem to power the controller (Atmega 328). Unless I use a reference zener and hang this from the positive, with a drain resistor to negative and hang the controller on the positive side: POS - CATH Zener - ANODE Zener (connected to negative of the controller) - DRAIN Resistor. The Zener current plus the controller current pass through the drain resistor to GND. Workable?

Click to expand...

You can power both the controller and the current monitor amplifier (whatever type) from a supply line, say 5V, derived from the PSU main supply. A Zener diode, as you describe, could be used or even a three terminal voltage regulator.

(having just returned from a holiday in Spain, I am having trouble even figuring Ohm's law at the moment.)

spec

Hi Spec, a 7805 can have max 28V over its input, so that is out which is why I think about a Zener; and a reference Zener as fixzd voltage for the controller would make the AD conversion more accurate.

Instrumentation amplifier: do you have some suggestions?

Holiday in Spain: if only Ohm's law is at issue you had not enough sangria

earckens said:

Hi Spec, a 7805 can have max 28V over its input, so that is out

Click to expand...

It is quit true that a 7805 will not take the in/out voltage to do the the job, but a TL783 set to 5V will: https://www.ti.com/lit/ds/symlink/tl783.pdf

earckens said:

which is why I think about a Zener; and a reference Zener as fixed voltage for the controller would make the AD conversion more accurate.

Click to expand...

Using shunt Zeners as power supplies is a good approach: simple and good protection from transients, but the current capability is limited.

earckens said:

Instrumentation amplifier: do you have some suggestions?

Click to expand...

On reflection, perhaps an INA193 (see post #51) would be worth investigating. An 0.08 Ohm sense resistor should be about right giving an output, from the INA193, of (20 V/V * 3A * 0.08 Ohms) = 4.8 Volts at 3A PSU output current.

earckens said:

Holiday in Spain: if only Ohm's law is at issue you had not enough sangria

Click to expand...

We had a wonderful holiday: hotel was perfect and much too much sangria, but it is nice to be back to simple food and the real world.

spec

Hi Spec, thanks for the TL783 info, I did not know of its existence. In my application the current draw would be between 20mA and 100mA for the controller and a I2C LCD display, so that is acceptable/doable with a Zener; plus a LM4040 5.0 will give me a reference voltage usefull for an accurate A/D conversion.
You mention INA193, but since I would hang the controller and sensor module from the high side, dropping with a 5V Zener and in parallel over this Zener the controller and sensor, would I not be able to use a INA219? The advantage being that on Aliexpress I can buy the whole sensor module (high accuracy B-version) for less then €5 (**broken link removed** and being lazy by nature I would not have to build circuitry around a INA193 chip (no module for that one in China): what is your opinion?

earckens said:

Hi Spec, reading your text and subsequently again studying your post 22 (and 12) I discovered an anomaly between what I think the layout should be and what you drafted in your posts. Now I may be wrong, and that is why I would like to ask you if I may pm you the circuit that you had drafted in post 12 with what I think should be the correct layout.
I would appreciate tremendously you taking a look at my drawing and comment on the corrections that I think should be in place. They all are centered around the layout for +IN and -IN.

Click to expand...

Hi earckens,

I have analyzed the the circuit of the PSU with the hand amendments (18K and 82K resistors) that you inserted in your PM: Yes, you are right. I was forgetting that the inputs to the error amplifier in the 723 do not operate down to 0V. Well done. I will amend my circuit to correct accordingly.

The 56 Ohm resistor you added provides protection for the current limit transistor so I will include that too.

spec

earckens said:

Hi Spec, thanks for the TL783 info, I did not know of its existence. In my application the current draw would be between 20mA and 100mA for the controller and a I2C LCD display, so that is acceptable/doable with a Zener; plus a LM4040 5.0 will give me a reference voltage useful for an accurate A/D conversion.
You mention INA193, but since I would hang the controller and sensor module from the high side, dropping with a 5V Zener and in parallel over this Zener the controller and sensor, would I not be able to use a INA219? The advantage being that on Aliexpress I can buy the whole sensor module (high accuracy B-version) for less then €5 (**broken link removed** and being lazy by nature I would not have to build circuitry around a INA193 chip (no module for that one in China): what is your opinion?

Click to expand...

Hmm, an interesting approach but I think you would need a separate isolated power supply for the INA219 + Microcontroller + display if I understand what you are proposing correctly. Otherwise you would be potentially injecting 100ma plus into the output of the PSU. Can you post a schematic of what you propose?

It sounds like you will be using an Arduino Mega for the microcontroller: is that correct?

spec

One of**broken link removed** based on the ACS712 chip may solve your problem as the current sense terminals are isolated from the rest of the electronics. It works by measuring the magnetic field using a hall effect sensor close to the current path.

Les.

spec said:

Hmm, an interesting approach but I think you would need a separate isolated power supply for the INA219 + Microcontroller + display if I understand what you are proposing correctly. Otherwise you would be potentially injecting 100ma plus into the output of the PSU. Can you post a schematic of what you propose?

It sounds like you will be using an Arduino Mega for the microcontroller: is that correct?

spec

Click to expand...

Hi Spec, included is a concept drawing for the current and voltage measurement. Since I hang the controller (Atmega Pro Mini 16M 5V) from the positive side, its ground hangs 5 V lower, yet 44Vdc (rectifier bridge output) - 5Vdc = 39Vdc above the PSU ground. The PSU regulates between 0Vdc and 30Vdc, therefor the voltage feedback to the controller is between -39Vdc + 0Vdc = -39Vdc and -39Vdc + 30Vdc = -9Vdc. So I need to convert voltages between -39Vdc and -9Vdc to 0Vdc and +5Vdc.

For the current measurement I insert the INA219 in the positive supply line. Powering the INA 219 is straight from the controller supply (over the Zener, 5Vdc).

To Les Jones: I am currently experimenting with the ACS712 but I get less than satisfactory results: the best stable resolution I currently get is worse then 100mA. I am experimenting with filter capacitors on the chip, with averaging readings etc. but I am not impressed so far.