0
\$\begingroup\$

I wanted to ask if anyone had guidance on making a transmission gate (with body diodes, however) to use on the high side of an H bridge.

For context, I am making an H-bridge to drive a resistive load. Unfortunately, I must use pre-approved components and these do not include any high-side / floating gate drivers or boost circuitry.

It wasn't too bad to make a dead-time circuit and use high-side PFETs and low-side NFETs, but there's one inconvenience: the positive rail going to the H-bridge load can vary from 0V to 12V with a constant 12V available to use for the driving circuitry. Thus, the PFETs may not have a high enough voltage on their source to be biased. My solution was to make something similar to a transmission gate with a discrete NFET in parallel to the high-side PFETs. Thus, when the rail going to the H-bridge is too low for the PFET, it's low enough to allow the high-side NFET to be biased by 12V and vice-versa when the rail going to the H-bridge approaches 12V.

I have attached an example of my proposed circuit constructed in LTSpice. H-bridge with both high-side NFETs and PFETs

  • R1 is the load
  • H_CTRL0 and H_CTRL1 are the digital signals used to control the H-bridge. They are timed to include a dead-time
  • M1+M4 and M2+M3 are the "transmission gates" that I am referring to here

It appears to simulate okay, but I wanted to ask a few questions here if anyone could answer:

  • Are there any obvious bad characteristics of this strategy that I'm not considering?
  • I understand it's not exactly a transmission gate due to the body diodes of the discrete FETs, but I believe I only lose the bidirectionality, right?
  • I do notice some large current pulses (maybe 20% more than the load current) when the FETs turn off in the simulation. I had hoped my dead-time circuitry would account for it (not included in the schematic, I just used correctly timed pulses instead), but the addition of the high-side NFETs is when this behavior started. Any ideas what the cause is or how to mitigate?
  • It seems like I can lower these current pulses by lowering the series resistance of the control signals H_CTRL0 and H_CTRL1 if that gives any hints.

Thank you for your time!

\$\endgroup\$
4
  • \$\begingroup\$ Did you try it with lower values for R8 and R9? \$\endgroup\$ Commented Sep 16 at 2:32
  • \$\begingroup\$ Is switching speed or efficiency a concern? \$\endgroup\$ Commented Sep 16 at 4:21
  • \$\begingroup\$ @Jens I did, didn't seem like it made a big difference. \$\endgroup\$ Commented Sep 17 at 0:57
  • \$\begingroup\$ @TimWilliams it's not, it will be operating in a single direction most of time \$\endgroup\$ Commented Sep 17 at 0:58

0

Reset to default

Your Answer

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

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.