0
\$\begingroup\$

I need to step up a 9V power supply (from a battery) to (at least) 45V, very small output current (less than 1mA, but it may be much less). My purpose is to drive a Russian rod vacuum tube, which needs 45V on its screen grid, and works best with about 60V on its anode (but the latter request may be reduced a lot). The current drawn by the screen grid is negligible, and the anodic current won't be more than 1mA.

I am thinking about a charge pump voltage multiplier; Maxim has many of them, but I have not been able to locate an IC suitable to multiply the input voltage by a factor greater than 2. I do not want to use boost converters, since I do not want to introduce an inductor in my circuit.

A possible solution would be that of connecting two charge pumps in cascade, but I wonder whether someone has a better and simpler idea.

\$\endgroup\$
7
  • \$\begingroup\$ Why not an inductor? A boost converter is the ideal choice. \$\endgroup\$
    – winny
    Commented Jan 14, 2017 at 22:49
  • 1
    \$\begingroup\$ Charge pumps are lossy due to ESR and 45mW out may need at least 200mW input with many stages whereas a boost regulator can be easily>90% efficient and better regulated with load current. \$\endgroup\$ Commented Jan 14, 2017 at 23:25
  • \$\begingroup\$ Your requirements are conflicting. Hidden in a comment (new information should really be added to the question) you say the reason for avoiding inductors is cost. That might matter in a high volume design. But then you say you want thru hole, which clearly says this is not a serious high volume design. This very much smells like a hobbyist fixating on one or two parameters and ignoring the bigger picture. If you're building one of these, then how is $2 in parts cost relevant? Also by insisting on archaic thru hole, you severely limit the choices and make everything harder to solder. \$\endgroup\$ Commented Jan 15, 2017 at 14:05
  • \$\begingroup\$ @OlinLathrop Yes, it is a semi-hobbyist design. Not intended for mass production. I understand I limit the choices, but I do not understand how everything be harder to solder---how could I solder SMD devices without the proper equipment (which I do not own)? \$\endgroup\$
    – Enrico
    Commented Jan 15, 2017 at 15:17
  • 1
    \$\begingroup\$ If you only have a soldering iron, you should avoid leadless packages like QFN or BGA. However, that still leaves many many parts available to you. Being able to use transistors in SOT-23 packages, microcontrollers in SOIC-28, TQFP-44, etc, is very useful. Many modern parts simply aren't available in thru hole because there is no market for such packages. \$\endgroup\$ Commented Jan 15, 2017 at 17:16

2 Answers 2

2
\$\begingroup\$

A basic capacitive charge pump can only double the voltage, minus voltage drops in diodes. You would need at least 3 stages of doubling. Other types of charge pumps add the input voltage (again, minus a bit) to the output each stagle. For such a charge pump you need at least 6 stages, since 45 is exactly 9 x 5, so 5 stages can't get there due to some inevitable loss.

So what's the solution? A boost converter. That uses a inductor to boost the voltage. You say you don't want to use a inductor. However, you provide no justification for that, so this is still a valid answer. We are not here to indulge your religious hangups.

A flyback converter uses a transformer. I don't know if such a beast falls into the catagory of evil inductors in your belief system. The sacred scrolls can be interpreted either way. In any case, a transformer-based power supply, whether in flyback or forward mode, can do this in a single stage, just like a inductor-based boost converter.

\$\endgroup\$
4
  • \$\begingroup\$ No religion here. I'd just like to save room and money, that's the reason for leaving out an inductor. But anyway, you convinced me that a boost converter is better, so I will convert my religious belief to inductors. Now the question turns to: are there around IC's suitable to my needs, I mean, suitable to very low output current, maybe without an external MOSFET? (I would add the request that the package be a through-hole device, not a SMD one.) But maybe now I am moving a bit off-topic and requesting info about specific products, I don't know. \$\endgroup\$
    – Enrico
    Commented Jan 14, 2017 at 23:46
  • \$\begingroup\$ @Enrico check out Linear Technology ICs \$\endgroup\$
    – crowie
    Commented Jan 15, 2017 at 2:53
  • \$\begingroup\$ @crowie The LT1172 seems to me a good choice. It is available in 8-pin PDIP. But the available application notes list quite complicated circuits. I'd need some help to simplify them and focus on my needs. \$\endgroup\$
    – Enrico
    Commented Jan 15, 2017 at 10:52
  • \$\begingroup\$ Since the output voltage of charge pump circuit using SMPS drops when current is drawn, what will happen when we have multiple stages of them in series? \$\endgroup\$
    – quantum231
    Commented Apr 15, 2023 at 3:37
1
\$\begingroup\$

Voltage multipliers will work out to 10* and 12* without any problem. But you need a reasonable high frequency (several kHz) and as noted you will end up using significantly more power to create your 45 V supply than you would using an inductor. If you need regulated 45 V then you may need to go to 6* or 7* because of the voltage drops over the diodes.

You could start here or here for some design guidelines and information. I've seen simple NE555 designs used out to 4*, I'm sure you could make one work at 5*.

If you want to design a boost regulated supply you can do no better than TI's Webench. This application can lead you through a design almost step by step or provide you with a starting point to further develop a solution. I put 9 V --> 45 V @ 100mA and it pointed out the LM2588 which is very much overkill, but it may give you some pointers.

\$\endgroup\$
1
  • \$\begingroup\$ Thanks for the useful references---anyway, you all have convinced me that a boost converter, with an inductor, would be better. But I see there are many around: could you please provide me with a good starting point for such a design? \$\endgroup\$
    – Enrico
    Commented Jan 14, 2017 at 23:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.