15V Low power DC-DC converter for HV supply (Hamamatsu C4900)

[luxLumen]

For https://imgur.com/RM2cYKF,SpTmMJl#0

Hello,
I need to convert 2*18650 Li-Ion batteries which are connected in series to 15V to power a C4900. (I could take 3*18650 in series if 2 are not possible or efficient to convert.) The spec says 15V ± 1V @ 14mA no load and 90mA when full load. It is for a portable application so the converter should be efficient and low power as possible, maybe also low noise / low output ripple, if that makes even sense. Which ICs converters would you recommend?

edit
Or should I take a ready-to-use DC-DC step up converter from ebay? There are devices like these which also mention that they replace the LM2577 with a XL6009 and filter out high frequency noise (see picture). These are for much more power devices than the ~2W that I would need and so they are probably less low power.

[Paul Price]

LMR62014

[luxLumen]

Thanks. As in my edit, what about devices from ebay which already do have the outside components soldered, are they suitable? Which of those "dc-dc step up converters"-searches are suitable for my purpose? I only need a few watts max output.

[Paul Price]

As far as EBay goes, I quickly see that the modules offered appear be larger, heavier, less efficient and use a much higher quiescent current. Reliability will vary with Chinese vendor.  The single IC DIY approach uses 9 total components, allows use of just one or optionally two Li-Ion batteries for greater efficiency.

1pcs MT3608 Step Up Power Apply Booster Module DC-DC 2V-24V 2A NEW
looks best 2 me on ebay.

[luxLumen]

Thanks, Paul. It's only after my initial question that I thought of buying already soldered small SMD PCBs, otherwise I'd probably straight away prioritised my question regarding these ready-to-use devices. They are probably less power efficient for my purpose (need 2 watts) but also probably less expensive and time consuming because I don't have a DIY PCB and the other 9 or more components here. So I'm going to choose something from ebay, maybe the one that you suggested but I'll compare it to a few others as well.

Anyone feel free to reply.

[mij59]

Current rating of power modules on ebay are mostly way to high, for save operation divide by 2..3
For what is worth,  step up module with the XL6009 input voltage 7,2V output voltage 15V.
Input current (mA)/output current (mA)
15/0
280/100
 

[luxLumen]

mij59, do you mean the efficiency? That looks not very high at all indeed. 280/100 -> 35.7% efficiency, that's bad. I guess it's better at higher loads but this is not what I need. The suggested LMR62014 appears at be better at 70-80% efficiency at the 15mA-100mA loads that I need (see 'LOAD (mA)' specs on page 6). Do you know any already soldered dc-dc converters that would fit better my needs?

blueskull, my was almost 3 times cheaper including shipping. I wouldn't pay 120$ for it indeed. I choose to get this for my PMTs (=photomultipliers) because it's designed for them, by the same manufacturer. It has protection features, is low ripple/noise and saves me a lot of time when I would have bought the individual parts myself and done the work. Of course this has its disadvantages as it's more expensive, less repairable and I learn less. But I wanted this to get working first. If you have schematics for building such a PMT supply, maybe as a kit to buy from somewhere, you may please post it, I'm certainly going to build it because as you can see in the picture there are 3 PMTs, so I need a HV supply for the other 2.

[luxLumen]

Here is a test of a XL6009 based DC to DC Boost Converter Module
The usual number in ebay descriptions says "Output Ripple:50mV". I think this is rather high (like 3-5 times less would be nice, PMTs need low ripple).

What about a precision low ripple reference voltage module? Are they designed for a 100mA load? From the name of it they seem not to be designed for a load but just measuring with a high ohm DMM?
Or what about modifying the XL6009 ebay module to reduce the ripple? Maybe bigger caps? I'm ok with the power consumption for the moment.

[Delta]

mij59, do you mean the efficiency? That looks not very high at all indeed. 280/100 -> 35.7% efficiency, that's bad.

Erm, nope. Try again mate!

input 280mA, but AT 7.2 VOLTS! (2016mW)
output 100mA, but AT 15 VOLTS! (1500mW)
so you're talking around 75% efficiency 😉

[luxLumen]

Thanks, Delta.

75% isn't too bad but the noise and ripple of a switching based supply isn't great for my application I found out. I also don't know how to reduce it (as already mentioned, maybe a filter cap? coil? I don't know, I'm a beginner). Ebay description says it's 50mV (not specified input voltage, loads) but with higher input voltages the noise is higher, and with lower loads it might be lower, so I guess the ripple may be lower than 50mV but could also be higher.
I saw vid (up to 1:18min is the relevant info) so I thought of using 4 Li-Ion and dropping it down using a LDO. But since 4*4.2V (when full charged) and 4*3.7 (when discharged) is in a range of 16.8V down to 14.8V I might even try using straight 4 Li-Ion in series, the noise and ripple should be very low, there would be no additional components and lower power consumption because no conversion. Of course the 16.8V are over the specified "+15V ± 1V", but since it's Made in Japan, I hope it doesn't break just because of the 0.8V extra, lol. The high voltage is variable with a 50k potentiometer from 0V to -1250V, so when the voltage drops after like a 0.5V I would may adjust it. Each PMT has a plateau characteristic but this kind of a 0.5V or even 1V drop which would translate into a 25V to 50V drop (@ typical 900V) which would fortunately hardly make any difference (at 2V drop it might make a difference but then the Li-Ions would be discharged anyway), so that's good. PMT handbook (p148 plateau characteristic, p155 HV suppy info).

[Delta]

Run five 3.7v Li Ion cells in series, giving a voltage range of 21v to 16.5 (assuming a max discharge down to 3.3v), then use an NPN linear regulator (not and LDO!) such as a 7815.

Fully charged, the reg will be dropping 6v @ 100ma = 600mW, load is 1500mW so efficiency is 60%, not great, but there will be next to zero ripple.
At 3.3v per cell, efficiency will be up to
1.5v x 100mA = 150mW
efficiency is now 90% 😃

As usual, there's a trade off between efficiency and ripple....

[luxLumen]

Thanks, I thought they had similar ripple, but you're saying that it has even less than a (low noise) LDO? And why not to use a LDO at all? If so, then this is definitely an option. I'd just need to compare them and see if a low noise LDO isn't enough, then the only 60% efficiency is a valid alternative.
From I see that because of the typical dropout voltage of 2V (and 0.5V for lower dropout voltage) indeed 5 Li-Ions would be needed. This all is a way better option than the first switching based high ripple supplies (because I don't know how to fix/modify the PCB).

So far:
0. Switch step up DC to DC converter from ebay, often bought by people, not an option because of 50mV (or more) ripple and nobody answered how to lower the ripple down to like 1-5mVpp.
1. Low Noise LDO: if the ripple is below or similar to that of a 7815 then the LDO would be better because of its efficiency.
2. 7815 or a 7815 with a lower dropout voltage, latter one is more expensive but this would translate into many hours extra.
3. 4 Li-Ion in series without any additional components and hoping that the extra 0.8V won't break the HV supply. Has the advantages of no additional components, probably even lower noise and longer running time (it's for a portable application so that's worth it).

A 7815 is worth it only if I know how much lower its noise vs a low noise LDO is. Any info?
*edit* I just saw the "Output Noise Voltage" for a 15V Voltage Regulator is "90µV". While for a Low Noise LDO it was 4.17 µVrms. Not sure what this all means. */edit*

I must say the 4th option is very tempting. Any reasons besides the 0.8V extra for not using this option?