HVPS in parallel

[Max DN]

Hello,

As mentioned in one of my previous posts, I'm building a nixie clock with Dalibor's RZ568M tubes.

The idea is to make the base as thin as possible and to achieve that I am using two power supplies, each with a transformer that is about 6mm tall.

If I build a clock with 6 tubes and 2 colons, I can allocate 3 tubes + 1 colon to each power supply and the brightness is nicely balanced.

If I build a clock with 4 tubes and 1 colon divider, then I am not sure if I should allocate 2 tubes + 1 colon to a power supply and 2 tubes only to the other power supply. If I do so, the brightness in the 2 tubes only will be different than in the 2 tubes connected to the colon. I could use PWM to adjust brightness but let's assume that I just do direct drive with no PWM.

I thought, is it possible to put two HVPS in parallel as you'd do with the batteries to increase overall current for a higher load? I don't think that's possible with DC/DC boost converters as it messes up the voltage on the top resistor of the voltage divider/feedback loop. I can always try and experiment but wanted to hear any suggestions before frying two power supplies.

I have built a larger power supply that can take all 6 Nixies etc but the transformer is much taller of course, so I am not using that one.

I could add a third mini power supply just for the colon, but it seems overkill to have 3 power supplies!

Any thoughts?

Thanks, as always.

Max

[Moses]

If the voltage from both supplies are about the same and the resistors on the tubes are the same.. the brightness will be about the same. I don't see how there would be a difference.

Regards,

-Moses

[Max Di Noi]

I thought so too, but I’m thinking that the regulated HV pulses from one power supply may impact the feedback resistor(s) on the other power supply, potentially messing up the feedback loops on both sides.

I could add a HV diode such that current doesn’t flow from one power supply into the top resistor of the feedback divider from the other power supply, if it makes any difference at all (drop out voltage is tiny anyway).

--
You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neonixie-l+...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/28dc3079-a49a-4fdb-aa23-03ecaaf82037n%40googlegroups.com.

[Kevin A.]

There are a few issues with paralleling supplies that are not explicitly designed for that mode of operation.

Imagine you have two voltage sources in parallel. The smallest difference between the two would cause current to flow from the higher source to the lower source. Depending on the control loop parameters of the DC-DC converter, this could cause it to shut down in an overcurrent condition or oscillate. 

Even if paralleling the supplies together does not cause issues without a load, once a load is applied, the voltage sag between the two independently controlled DC-DC converters will also differ. So again we have conditions for oscillation or regulation problems.

I would recommend using a single supply that meets the total system capacity, or keep the individual supplies isolated. If you want to try paralleling, there are a few schemes you can explore which might work better than direct paralleling. Here is se reading I found on the subject: https://powerctc.com/en/node/4577

--

[gregebert]

Even using diode-isolation between the supplies could still result in 1 supply driving most of the load.

Though wasteful, you can add series resistors (ballast resistors) from each supply, to help balance-out the loading. 

I think the best solution is to use a single supply that will provide the required current, plus some extra margin. If that isn't possible, split the loads so they are on independent supplies.

[Moses]

Going to clarify a but.. I was referring to using separate supplies for tubes, NOT paralleling them together. As others have said problems arise when you parallel two switch mode supplies.

Back to the comment, if you use one supply for one tube or two or ?? And another for colons or other tubes, doesn't matter the number. I really don't see a problem.

Regards,

-Moses

[Richard Scales]

On my boards for 6 x  RZ568M tubes I do exactly that - I use two HV supplies and have divided the HV supply line into two - one for the left and one for the right, one HV module for each.

If using 4 tubes and one colon - and you still want to share the load between two supplies - then why not  split the supply to the two neons in the colon - one to the left and one to the right?

If your colon has only one neon then are you concerned that the effect of lighting one neon will make a visible difference to the overall tube brightness on one side? I would have thought (but have not tested) that the difference would not be noticable (ready to be proven wrong!).

Assuming that is the case - then what effect would the difference between displaying an '1' or and '8' one one tube (different currents) have on the adjacent tube on the same HV Supply?

Put yet another way - is there a visible effect on any tube on any clock when a colon neon is illuminated? (there could be - I've just never noticed it - also I have a head full of a cold at the moment and all thinking is clouded!).

I don't know what HV modules you are thinking of using and although I've not done this myself - I did see a neat trick where someone was using Omnixie NCH8200HV modules which they recessed into a suitable hole in the PCB - assuming space permits - this allows you to trim the PCB thickness off the overall height.

- RIchard

[David Forbes]

I would hope that your power supplies have regulated voltage, correct? Therefore they should both be the same voltage, within the tolerance of the voltage determining components, if the supply design has good load regulation. In the worst case, provide a voltage trimmer on one supply so that you can adjust it to balance the brightness of both sections.

You may want to do that even in the case of equal loads, since a 1% difference in HV voltage will make about a 5% difference in cathode current due to the fact that the dropping resistor works over a small fraction of the HV voltage. 

On Fri, Aug 19, 2022, 3:12 PM Max DN <flata...@gmail.com> wrote:

Hello,

As mentioned in one of my previous posts, I'm building a nixie clock with Dalibor's RZ568M tubes.

[Max Di Noi]

Hi Richard, David, 

I'm using my design of the NCH8200HV, it's mostly its replica as I wanted to embed the design in my board and I'm not bothered about the 'not invented here' syndrome as I'm not selling commercially. 'Sinking' the HVPS into a pcb cutout is a good idea though, I had seen that too. 

I thought about allocating each neon of the colon to a different power supply but I'm using the colon made by Dalibor and it has only one anode and two cathodes, so...

I may add a separate power supply just for the colon and I have a trimmer resistor on each power supply (thanks for the suggestion David) to adjust brightness before doing anything via software. 

You are right that turning on '1' and '1' (say for Hours) on the same power supply may have different brightness than '4' and '8' on the other supply (say for Minutes). I'll be using different resistors for each digit as suggest by Dalibor, there could still be a small difference in brightness, but I'm not planning to solve for that. I don't think it's noticeable, if it is, I'll adjust from software, but there are many combinations to consider, so that's definitely overkill.

Of course I have and I can use one big power supply for everything but I'm trying to keep the final case as thin as possible.

--
You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neonixie-l+...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/CAPbqtvcAMk0OysoHzLyizAKCZBWG9zzoSxxe3JxSOdgCYZZDJQ%40mail.gmail.com.