High Power Mosfet Output Stage, More Power to your Track Options

[RichG]

Not completely ready on this but thought I would get the Topic started.


Part 1. Introduction. The Standard Powerbase and the Test Setup.


So now that the building works are complete I've moved into the new study (Laboratory?) and have room to move. Time to get back to work.



Here's the test setup. A modified Powerbase in the foreground, more on that later. In the background a 30 Amp PSU top right and the electronic load, bottom left.

Electronic load is DC only so it is fed with a Schottky Diode bridge, not very visible but at the end of the Yellow & Green wires, the Black and Red wires go to the load. This bridge is effectively the same as the bridge in the in-car controller, so the voltages measured as the same as the motor electronics will see after rectification.


So how does the Standard Powerbase perform under load?




Well as we all know not very well.


Input voltage in this and all the other tests is 13.5 Volts. With a light load (0.1 Amp) this voltage is 12.36 by the time it gets to the electronic load. This is actually fine as there are various known losses along the way. Connecting leads, protection diodes, rectification loss etc. What matters is how many more volts we loose when we increase the load.

So looking at the second column you can see that by the time we get to 3.5 Amps the voltage is down to 5.43 Volts. This is pretty useless so I did not bother increasing the load any further. It's also the reason we have set Overload in PB-Pro to 3.5 Amps. The other problem is that the voltage reduction is not linear, it increases with load. Initially there is a 0.8 Volt change for 0.5 Amp increase in load however there is a 1.5 Volt change at 3.5 Amps.


There are several reasons for the losses, I will go into more detail when I show you the figures for the improved output stage.

The biggest ones are the losses due to RDS on, the resistance of the mosfets. For the ones in the 1.5 Powerbase this is 0.07 Ohm plus 0.175 ohm. These being for the P & N channel mosfets. Total is 0.245 Ohms. So roughly 1/4 Ohm which at 3.5 Amps gives us 0.875 Volts, so significant but not the main culpret as we are loosing a total of 6.93 Volts.

Another smallish contributer are the chokes between the mosfets and the track. If these are removed you get the third column. So another 0.82 Of a Volt. I have removed these for all subsequent tests.

So that's about 1.7 Volts accounted for but still over 5 volts to go. I will explain the other losses when I talk about the new output stage but accept for now that the lions share of this losss is due to the mosfets being poorly driven, both during the switching time and when switched on. This then of course creates a double wammy, increasing lost volts with current and heat.


So to be continued with a look at a new power stage and some new figures.



Rich

[Ian B]

WOW Rich,
Good to see you have settled into your new Lab. I know how long you have been waiting

I'll be following the results too

[300SLR]

Perhaps I've misunderstood - please tell me if this is this the power going to the track to drive the cars and lane change mechanisms.

[RichG]

Part 2 First look at the New Power Stage.

What we have here are 4 N Channel mosfets with heat sinks. A HIP4081 mosfet driver chip and a few discrete components to make it all work. I will put up a circuit diagram in a later post. Please also note that this is just a prototype to make some measurements on, final version not really thought about yet, but I have in mind that it might go in a small diecast box that would double up as a heatsink. You can see one in the background in the first post.



The mosfet driver chip uses the same signals that we pick off for the master slave mod, the internal stage is completely disconnected. This of course means that you can use this in a 1.2 Powerbase. The driver chip ensures that the mosfets are fully driven and also enables us to use N-Channel mosfets in all 4 positions. These have a lower RDS On & are available as protected devices.


Have tried two devices, the first one an IRF 530N, this is a conventional mosfet rated at 70W and has an RDS ON of .09 ohm.
The next column shows the VNP20N07, this is a protected mosfet that amongst other things shuts down if it gets too hot. It's rated at 83 Watts, and has an RDS On of .05 Ohm.

So here are some figures.




The protected device performs better, largely because of it's lower RDS ON, so with that and the benefits of Autoprotection I will use that column to explain some of the numbers.


So these numbers are more like it, but where do the volts go?

First thing to note is that the volt drop for each 0.5 Amp increase in current are linear and are about 0.2 Volts.

Total volt drop between 0.1Amp & 5 Amps is 1.9 Volts.

Some of the losses can actually be improved on, this is what I have measured and calculated. All of the individual numbers look small but this is a usefull reminder that they add up.



1) PSU regulation 0.16 Volts. It's a big PSU but not perfect

2) Connecting lead to the output stage. 0.23 Volts. I have run a separate lead, the power does not come from P1 & P2, but needs to be bigger.

3) Sense resistor 0.25 Volts. Needed to detect when there is an overload. I have left the standard .05 ohm resistor, it could be reduced in value.

4) Output lead. 0.24 Volts. This could be improved but in practice could be worse than this as this is equivilant to the track.

5) RDS On for two mosfets 0.5 Volts

So that's a total of 1.38 Volts. Leaves 0.52 Volt unnacounted for, this will be in the switching, change in vf of the schottky diodes, other losses, etc.


I have wound this quickly up to 8 Amps and the volt drop was still linear. I also did a test at 5 Amps with no heatsinks, after 3 minutes the case temperature was over 100 deg C and the thermal protection cut in. With these heatsinks it stabliised at 50 deg C with a constant 5 Amps being taken.


So a lot more work to do and I have yet to try it on the track, but I feel this is a step in the right direction. I do not have any numbers on MS or MSS so cannot make any comparisons.


To be continued.....



Rich

300SLR I do not understand the question.

[awallace]

OMG - RichG is taking more stuff to bits - is nothing sacred?! I am really looking forward to this one, and have just withdrawn a bid on a v1.5 powerbase on fleabay in the excitement - that is one hell of a study youv'e got there by the way!

[Budaholicguy]

I know I'm probably jumping the gun a bit but is this intended for wires to be soldered into the PB circuit board as the MS mod or will it pick up from the 'aux out' socket and just involve cutting the PB supply wires to the track and replace with wires from the little silver box??? just curious, the less 'cut and shut' needed to the PB the better for me!!!

Great work (as always)

Cheers

Bud

[RikoRocket]

He'll be working a miracle if he can get the power out through the aux port! Sorry Bud I'll take a sweeping guess, its gonna take drive off the same places as MS.
Riko

[RichG]

Riko is right, although it was in the text of my post. Picking off the signals will I am afraid be a bit like doing a master slave mod.

In addition there will be 2 larger wires to return the power. My outline plan is 3 small wires, 2 signal, 1 ground to an RJ11 style plug and two larger wires to return the power. These will exit the powerbase and connect into the die cast Box. The larger wires into a terminal strip along with a direct feed from the PSU, making 4 in total. This will facilitate easy changing of the Power stage for testing etc.


If all goes well I have in mind, for those that don't want to make their own, that we could offer an upgrade of 1.2,3,4 PB's to PB-Pro and the uprated Power stage at the same time. However a lot more testing yet and a circuit board to be layed out etc.


Rich

[Budaholicguy]

'He'll be working a miracle if he can get the power out through the aux port!'

Riko, the impossible seems to get done right away, miracles just take a little longer!

[Reiner]

Hello Rich
I read the post for more power with a lot of interest. If I understand the change right to more power that you will offer in the future an upgrade for the PB Pro? Sorry for my english.

regards Reiner
Germany

[RichG]

Hi Reiner

It's early days, I am sharing the info on this well ahead of any "productionised" upgrade. However if all goes well I propose making this upgrade available to people with 1.2 to 1.4 Powerbase's to be done at the same time as a PB-Pro upgrade. I suspect there are a lot of older PB's around and this would give them a new lease of life.

Of course this could be done to a standard Powerbase but it fits well with the PB-Pro upgrade as we can set the overload level in software.


Rich

[GregR]

OK Rich, now I am excited .

I have been watching the PB Pro thread while I have been working overseas and as great as that sounds, what I am really after is consistent power and this sounds like the answer. MS is good, but I suspect this will be much better.

I am finally back home in a few days and I was going to change my MS into an MSS but now I might just wait Those 2 brand new V1.6 PB's might come in handy.

Well done

[chief32s]

Rich, I have a V1.2 and a V1.5 powerbase. I also have a 4carPB. I have been intending on PMing you regarding a PBpro upgrade, but have also been a little scared (and confused) of trying the MS mod.

Would this combined mod (PBpro and power) be available to both v1.2 and v1.5?
Apologies for my ignorance.

Keep up the good work.

P.S. I love the "lab". Reminds me of my father's attic.

[rolloverdj]

Rich,

Great work, looking at the results it definitely looks a solid step in the right direction.

Just a quick question, at this early stage does it look possible that you will be able to remove the mosfet PCB and replace it with a drop in replacement pcb with what your doing with no extra wires?

cheers
Phil

[RichG]

Morning Guys

A few answers to questions.

The mod will be equally applicable to all Powerbases. The reason I have mentioned the early ones, 1.2 to 1.4 is that we have chosen not to upgrade these to PB-Pro because of the weak output stage. Also I suspect there are a lot of the older ones around not being used, this would give them a new lease of life.

I have yet to work out the detail of how the mod will be fitted to the Powerbase, however I do not expect it to be a replacement for the 1.5 Mosfet daughter board. The size would be too small and fixing the heatsinks would be difficult.

My plan for the first prototypes is as described in the earlier post. The Powerbase will have it's mosfets removed and drive disconected. The drive signals will be brought out onto an RJ11 plug that will plug into one side of a small diecast box. The Power feed will be two wires from the Powerbase that connect to screw terminals on the other side of the box along with screw terminals for a direct feed from the PSU.

This will enable quick changes to be made to the Power Module without dissasembling the Powewbase. Once it's been tested in the field and proven to be reliable I will think about how best to fit it in a more permanant / attractive fashion. That said a plugable Power module is not such a bad idea and I may just stick with it.

There may even be different versions of the Power module. I am testing some very low RDS ON Mosfets this morning, 0.018 Ohms, these seem happy to run at 5 Amps continuous without heatsinks, in free air the case temperature has stabilised at 60 deg C. Another posibility is to double up on the mosfets which would of course halve the RDS On.


Rich