Author Topic: Terrible Idea Boost Converter To Be Tested On Gocart  (Read 3610 times)

0 Members and 1 Guest are viewing this topic.

Offline james_s

  • Super Contributor
  • ***
  • Posts: 21611
  • Country: us
Re: Terrible Idea Boost Converter To Be Tested On Gocart
« Reply #25 on: January 03, 2020, 06:35:55 am »
For a relatively cheap and quick solution I would suggest getting an inverter (one with 120VAC output will be ~170VDC) and either putting a rectifier on the output or tapping into its internal HVDC bus.

You can buy just the 12V->170VDC part from China, at least I've seen them available in the past.

Personally I'd be looking at an array of RC LiPo packs to get >100V or a surplus EV battery module. Salvaged Tesla batteries are surprisingly affordable for what you get.
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22436
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Terrible Idea Boost Converter To Be Tested On Gocart
« Reply #26 on: January 03, 2020, 07:20:07 am »
Consider what happens if there is a small say 1% duty mismatch between channels driven at the same PWM; consider further if they have negligible resistance (in the inductor, switch and diode) (which they do); and consider if there's say a 10% error in inductance.  What does this do to current flow in the different channels?  How can you predict this effect?  How can you prevent transistors from exploding?

Well, you can't, which is why we put current sensors on them to be sure. ;D

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline pepelevamp

  • Regular Contributor
  • *
  • Posts: 101
  • Country: nz
Re: Terrible Idea Boost Converter To Be Tested On Gocart
« Reply #27 on: January 03, 2020, 08:17:21 am »
get one of those spinning wheels that one motor turns while a second one gets spun & generates a voltage.
why bother with fast switching mosfets and silicon when you can add an extra wheel to your go-kart and use cogs to convert your power.  :box:  :) ;D :D

 

Offline i_am_fubarTopic starter

  • Regular Contributor
  • *
  • Posts: 53
  • Country: gb
Re: Terrible Idea Boost Converter To Be Tested On Gocart
« Reply #28 on: January 03, 2020, 08:44:14 am »
Hmmmm, good point. There is definitely a risk here.

I would argue that I can prevent the transistors from exploding. But it would be at a significant performance limit. Assuming that all inductors are actually at their -20% limit, I can configure my maximum on time to ensure their current can't rise above a safe value (say 50A). Yes, this means that if the inductor is actually at +20% tolerance, then I'm running nowhere near peak power throughput for the power components, but at the same time, it does mean nothing will go bang (in normal operation).

The mismatch is a bigger worry and I may end up with 1 of the 4 phases doing 50% of the conversion, but again, providing my timing is such that I'm not exceeding SOA for each device, I don't see it as a magic smoke risk.

Of cause, this is all my opinion as an engineer very new to high power, high speed switching.... So, please keep punching holes :)

Plus, if I up my input voltage to 24v at half the current, then the performance degradation of playing it safe should be offset.

Way to do efficient, low cost, design :) (sarcasm)

On the bright side, I think I finally grasped the physics of gate ringing last night and how the parasitics come in to play, why snubbers work, and why slowing the gate reduces the effect. So, as a learning exercise, this is working :)

 

Offline i_am_fubarTopic starter

  • Regular Contributor
  • *
  • Posts: 53
  • Country: gb
Re: Terrible Idea Boost Converter To Be Tested On Gocart
« Reply #29 on: January 03, 2020, 08:46:48 am »
Also, I have a bus bar design that I'm going to use for feeding the battery supply into each channel. I could integrate a current sensor into each branch of that to get independent monitoring.

Onward an upward, more bits coming in today.
 

Offline pepelevamp

  • Regular Contributor
  • *
  • Posts: 101
  • Country: nz
Re: Terrible Idea Boost Converter To Be Tested On Gocart
« Reply #30 on: January 03, 2020, 08:57:15 am »
this sounds like a real cool project. i am looking forward to seeing footage of you sitting in a go kart moving along off a 12v battery. maybe race a friend riding a power drill.
 

Offline jbb

  • Super Contributor
  • ***
  • Posts: 1265
  • Country: nz
Re: Terrible Idea Boost Converter To Be Tested On Gocart
« Reply #31 on: January 03, 2020, 06:29:59 pm »
You need a per-channel sense.  Think about current sharing, let alone compensation.

...Or do you think that the channel characteristics will vary that much?...

Yes, you almost always need it per-channel.  This is because a small change in switch duty cycle translates to a small change in the DC operating point of the inductor.  Because your MOSFETs and inductors will be chosen to have low DC resistance, you will get a large current imbalance.

However, there is one way to get around that.  It will likely compromise the efficiency of your converter. Go to Discontinuous Conduction Mode (DCM).

Advantages:
  • MOSFETs turn on near zero current.  No diode reverse recovery loss in synchronous rectifier diodes (which is good, because 200V MOSFETs have pretty crap body diodes)
  • Inductor currents drop to zero each switching cycle, which reduces the system order and can make compensation design easier
  • Current sharing will be largely - and stably! - determined by inductor tolerance
  • Inductor values will be lower

Disadvantages are significant:
  • Increased RMS currents in the converter mean higher conduction losses
  • Increased current ripple will increase inductor core loss.  Inductors may end up being quite large and difficult
  • Somewhat more difficult to control synchronous rectifiers
  • Inductor currents drop to zero and will ring a bit (inductance + MOSFET capacitance = LC resonant tank at a few MHz).  This could make a lot of radio noise, and may require snubbers to damp it away.


Here's a question: do you really need a well controlled output voltage?  I assume you will have some kind of motor speed controller after this stage (so that you can control the motor speed down to zero)?

Getting a 1:10 step up out of a boost converter is a bit of an ask anyway, so why not use a fixed-ratio step-up converter (e.g. unregulated full bridge transformer)?

 

Offline amyk

  • Super Contributor
  • ***
  • Posts: 8526
Re: Terrible Idea Boost Converter To Be Tested On Gocart
« Reply #32 on: January 04, 2020, 04:29:57 am »
Re. Inverter. Considered it, but the goal is to get some experience designing high power switching systems. So not really applicable :(
You can take the inverter apart and learn from its design too.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf