Which voltage in your application?

It’s good to know what voltage range people want to use ODrive for in their application. Right now the voltage rating is for 24V, but it’s fairly easy to change. There is a current handling penalty to increase it though, so ti might be useful to offer two version.
To help us decide between the different options, please answer the poll about what supply voltage would be the most ideal in your application. You may check multiple options if they apply to you.


What supply voltage is good for your application(s)?

  • around 12V
  • around 24V
  • around 36V
  • around 48V
  • >48V

0 voters

‘low voltage’ is legally defined as <50v, so anything above that requires
additional certification/safety measures.

@51v you are required to handle all the wiring the same as line voltage in your house (110v/220v)

Other than that, if it’s just a couple bucks, make it 48v safe (say 60v drivers
to have them survive spikes) and at that point it should be able to handle lower
voltages as well.

One issue are the bus capacitors: for a given capacitor technology the capacitance goes down with voltage squared, everything else being the same (especially volume).

Also the increase in price for upgrading the MOSFETs to 60V, leaving the Rds_on the same (which is required to not sacrifice current handling), would increase the BOM cost by $24, which is quite significant, especially if the gross margin from ODrive Robotics selling the board is applied as a percentage.

So we need to solve a 2-dimensional problem:

  • Offer one board (where we strike the right compromise), or offer two boards?
  • If we offer two boards, how do we strike the balance between current handling and price on the high voltage board?

Current ODrive board costs about the same as VESC board (depending on the supplier of course). But ODrive can handle two motors 24V/100A and VESC can handle one motor 50V/50A (I believe). Having ODrive board good for 50V/50A for about the same price as now would be a very competitive offer. Please note I don’t know how the pricing of the components work - I am just guessing it could be similar since the total power is similar. I will not need 5kW board in the near future - 2.5kW is quite enough :wink:.

If you decide to make 50V/50A board, please put me on the waiting list.

1 Like

Only supply voltage may not exceed 50V AC. As the ODrive is actually supplied with DC, it may not exceed 120V DC to still be covered by extra low voltage regulations.

The 50V AC is an RMS value, thus a peak AC voltage of 70.7V is still within those limits.

[Off Topic] for costs sake, i would get rid of that DC-DC converter. Converting grid power to 12V and then back to 24V or 48V lacks of sense. Better use an ordinary 24V powersupply. [/OT]

How much of a current handling penalty are you expecting as the voltage increases? 50% penalty at 48V, or less? I’m basically interested in the maximum power output that this system is capable of, and I would think that would increase as the voltage increases. The motors I am running (SK3-6374) are only rated for 44V windings so there’s not as much benefit in going beyond 48V for me. However, motor heating should be less at higher voltage and lower current, or so my EE professors would have me believe…

I’m a little leery of going on a significant design revision to push the voltage though, when nobody has tested the limits of the current 24V version. If it could sustain 1 kW or even 500W continuously that would be sufficient for my application. Do we have any estimates of what the continuous power output will look like? I’d love to see some higher power tests for both peak and continuous limits.

48V is standard for DC electric house buses as well electric cars. Also, being able to send more power through thinner cables seems like the best option to me. Over 50V is regulated differently which is probably why it’s avoided in new applications.

Yes you make a very good point about testing the limits of the current ODrive before going too far in new directions. To that end I just ordered a FLIR C2. Once it arrives (and I have some time heh) I’ll set up a stress test and we can work towards having a good understanding of the continuous spec of the current ODrive.

Whatever you do don’t make two versions! Save yourself some time and money, there are always people wanting higher voltage or higher current but that will still be true even if you make three versions. Now for my opinion on voltage:

I think it’s best not to look at what voltages people are using right now, but instead at how we can match available hardware. Let’s look at it this way: what is the KV range of the typical brushless motors we may be using, and what are the typical rpm we may be looking for? the KV is often around 170-300, and lets’s say for servo applications that we are usually wanting 2000-6000 rpm peak (I’m thinking direct driven ballscrews or pulleys for belts would be the most common uses). That would translate to 12-24v. 12-36v seems to be the most easily available power supplies as well.

So my vote is for something around 36 volts, with higher current abilities and cheaper capacitors. maybe even go 24 volts if makes the product cheaper, but since a lot of people are wishing for even higher voltages maybe 36 volts would be the best compromise. Having a voltage that completely mismatches the motor is a bit of a waste. Obviously we can rewind the motors, at least if they are outrunners, but generally I’d say it’s easier to just get a new power supply.

Great decision to have 3 motors per board by the way!

Edit: After looking at the list of recommended motors I’m thinking the current 24v may me best? You can always find a higher KV motor if you want more speed, but the lower KV motors are expensive and difficult to find.

I voted for >48 volts moslty because the prototypes I’ve developed for my system use off the shelf industrial systems running at higher voltages. The first one was 180 VDC, and my current setup is 75 VDC. I like David’s approach of starting from what’s availlable though. Other than lower currents and the associated losses, what other reasons would servo manufacturers have for using such high voltages given the constraints (capacitance, components costs, etc?) Since my ultimate goal is to build a consumer grade product, I am sensitive to cost as well as heat build up, so I’ll have to strike a balance.

Also, if cost is a concern can the ODrive be made modular, where the higher cost parts are place on shields like this:


That is installed on this:


Just a thought, may not be feasible though…

Can you develop this statement a bit more? When placing capacitors in series the maximum voltage will grow linearly with the number of capacitors, while keeping the capacitance constant.

[EDIT] I found the flaw in my reasoning. The capacitance will drop linearly when placing capactitors in series. Sorry for the noise.

1 Like

This is not true. The capacitance will be halved. See an explanation here.

One way to think of it is: the amount of energy stored per unit volume of a capacitor technology is roughly constant. The energy stored in a capacitor is 0.5 C V^2, as is explained here.

Thank you. But I still have a hard time understanding why we need to store more energy in our capacitors in a 50V/50A setup than a 25V/100A setup, with the same peak output power?

Think of it the other way around:
Same board layout -> Same number of caps -> Same volume -> Same energy.
Then if you solve E = 0.5 C V^2, you see that for the same energy, if you double the voltage, the capacitance goes down by factor 4. This means 4 times more voltage ripple on the bus per unit current, and importantly not 2 times. This means that the capacitors may be a limiting factor in scaling up voltage and sacrificing current.

Unless you specify the voltage ripple as a percentage of the bus voltage (rather than absolue), in which case it scales “nicely” with the same-output-power current/voltage tradeoff.

1 Like

but the voltage limits on capacitors are not volume based, it’s based on the
dielectric used. The higher voltage caps will be more expensive, but not
necessarily larger.

now, there are natural breakpoints due to the capacitor voltage ratings
available, I’m not familiar enough to know where they are.

Forgive me if this is obvious to everyone but me, but why do we want to specify the voltage ripple in absolute terms and not be happy with a percentage of the bus voltage?

I’m not an expert on interference, but I assume the noise that is produced is proportional to the absolute voltage ripple. It could make the product unreliable, and make it harder to comply with emission standards.

I think you hit the voltage perfect with 24V but making it compatible for 48V seems like it would be favored by some. Just from experience, I would not make two boards as it causes a lot more work and headache. Just my opinion.


1 Like

+1 camrbuss, the DRV8301 does not seem to handle voltage spakes >48V so well, therfore a bus voltage of 48V seems to make the unit less reliable. Swapping the DRV driver to a different part does not make much sense at this state IMO