Honda Insight Forum banner

1 - 20 of 202 Posts

·
Administrator
Joined
·
11,556 Posts
Discussion Starter #1 (Edited)
We have had successful LTO 60, 72, 84 cell conversions, and even my proposed 96 cells as feasible.

But can we go the other way and make it cheaper and easier to fit with only 48 cells?
4 blocks of 12 and a breeze to fit in the OEM pack space.

Instead of fooling the car (MCM/BCM) the voltage is lower than it actually is, we would need to fool it into thinking the voltage is higher than actual to get it into it's comfort zone.
We need to add a gain/multiplication factor to the various voltage inputs so if the voltage is 100V it tells the car it is 150V or whatever.

Can it be done reasonably easily and how? OK some random ramblings..

The standard car detects voltage at three points..

3) The 10x BCM voltage taps. (Pack voltage / 10)
2) The MCM high voltage input. (Full pack voltage)
1) The MCM 0-5V VPIN input. (0-5V)


1) We can use a PIC ADC to watch the 0-5V voltage coming out of the MDM VPIN output.
We process it and pass it thru with a multiplication/gain factor into the 8 bit 0-5V DAC output of the PIC and into the MCM VPIN input.
Fairly easy to do I suspect. Or some sort of non inverting op amp with adjustable gain..
Maybe be even a pull up resistor or two..

2) More difficult as we have to reverse engineer the MCM HV input circuit and increase it's gain or sensitivity by a suitable amount.
I did do work on this with my supercaps setup as I supplied it with a fixed voltage IIRC derived from a LV supply.
I changed (reduced) a few resistor values on the MCM board. Again probably doable.

3) If we use the BCM replacer it's easy, we just add a multiplication/gain factor to the software.
Using the stock BCM would be more difficult as we have to provide ten independent voltages with a suitable gain. Probably doable..

Other considerations.

How low can the stock DC-DC actually operate.
IIRC it was down to 80V or less before it shut down, so that's probably OK.
It will draw more current for the same load at lower input voltages.
So if the input voltage was 160v and it draws 5A then at 80V it would draw around 10A for the same output roughly.

If you have 48 cells instead of 96 you only have half the wh/ah capacity. ~2.4kwh instead of 4.8kwh

120V would be 48 x 2.5V and fully charged.
96V would be 48 x 2V and empty.

Anyway if i can get it to work with 48 cells then I have enough LTO here for 2 cars!!
A lucky UK owner might be able to be an LTO guinea pig.
 

·
Super Moderator
Joined
·
8,117 Posts
I think you are onto something good. One of the BIG advantages of the configuration would be mechanical. Two blocks could be placed fore/aft and fit "fairly" comfortably in the IPU on a square aluminum mounting rack, ala Natalia.

Obviously, there are several electrical considerations as you say and I'm not qualified to comment much, but I too will be watching with interest. It is a great idea if it can be made to work. Could be virtually drop in from some production shop.:)
 

·
Administrator
Joined
·
11,556 Posts
Discussion Starter #4
OK starting with the VPIN it's probably easiest to use an OPAMP.

This PROTEUS simulation shows a simple OPAMP adding 50% to our reported voltage.

We could make the increase factor variable by adding variable pots for R1 & R2

84589
 

·
Administrator
Joined
·
11,556 Posts
Discussion Starter #5 (Edited)
Moving on to the MCM HV input.
The partial schematic I managed to glean from looking at the circuit on the MCM PCB.

Can some others have a close look at an MCM and see what they can add.
Check my resistor capacitor values etc.
If we can simulate it then we can adjust the values more easily and accurately.

There were a couple of connections I could not trace.
I also post the link to my original investigations into this area.

Original Supercaps Thread HV MCM posts

EDIT Old schematic deleted.
 

·
Super Moderator
Joined
·
8,117 Posts
Not to be too distracting, but I did a bit of layout work for a 2 block configuration. There is a combination of 1.5x1" aluminum rectangular cross section tubing which forms a very simple 2 rail frame, since the blocks are mounted with tabs to the side. It is a take-off of Adria and Natalia's approach. If everything works out, they may want to offer kits for sale.

There is also an abundance of room, since the blocks get mounted "sideways."
 

·
Administrator
Joined
·
11,556 Posts
Discussion Starter #7
An easy way to fool the standard BCM voltage taps and the MCM HV input at the same time without modifying either of them would be use a voltage controlled 0-250V DC generator feeding a standard 0.1% tolerance 10k BCM fooler matrix and the MCM HV input.

Such a device does exist ... But it's expensive $170


Can we make our own cheap proportional high voltage doubler that uses the DC input from the pack?

If not a doubler one that generates +50% extra voltage to match our VPIN +50% opamp solution.

 

·
Registered
Joined
·
648 Posts
I’ve long since wondered if the IMA could be modified for very low voltages to operate more as a big starter.

given the IMA can make useful power through 5200rpms at 144volts one would think it could output useful power through 1300rpms @ 72 volts which is enough to start and nudge if the electronics could still function.

will watch this thread with interest
 

·
Administrator
Joined
·
11,556 Posts
Discussion Starter #9
OK here is a voltage doubler schematic.

84601


The input is our 48 cell battery (100v) in this example converted into a 400hz sine wave.

You will see from the various voltmeters scattered around the circuit doubles the battery voltage first to ~198V

(We lose a couple of volts in the diodes etc)

We then use a voltage divider (basically the BCM fooler with an extra pre divider resistor R0 = 30k) to bring the voltage down to where we want it, in this case battery voltage x 1.5

The 10 x 10k 0.1% resistor BCM fooler subdivides that 150V into ten equal taps for our BCM.

This would track the battery voltage and drive an unmodified BCM and MCM HV inputs :)

Now just need a sine wave driver at the front end..
 

·
Administrator
Joined
·
11,556 Posts
Discussion Starter #10
Driving the above might be achieved by an ICL8038 chip with op amp follower and a HV switching transistor.
 

Attachments

·
Super Moderator
Joined
·
8,117 Posts
OK here is a voltage doubler schematic.

View attachment 84601

The input is our 48 cell battery (100v) in this example converted into a 400hz sine wave.

You will see from the various voltmeters scattered around the circuit doubles the battery voltage first to ~198V

(We lose a couple of volts in the diodes etc)

We then use a voltage divider (basically the BCM fooler with an extra pre divider resistor R0 = 30k) to bring the voltage down to where we want it, in this case battery voltage x 1.5

The 10 x 10k 0.1% resistor BCM fooler subdivides that 150V into ten equal taps for our BCM.

This would track the battery voltage and drive an unmodified BCM and MCM HV inputs :)

Now just need a sine wave driver at the front end..
It looks pretty simple and I think the idea of an unmodified BCM and MCM could be very attractive. If the circuitry can be kept simple and inexpensive, the cost of 2 packs/block ($540 at InsightFest) is going to be a strong inducement for folks to chuck the NiMH "habit."

Do you think that the lower end of the battery voltage, about 96V by my calculations, will produce usable assist with an unmodified motor/generation.

I think I understand how the doubler part of the circuit works, but I need to find a ref link to study the waveforms a bit.
 

·
Administrator
Joined
·
11,556 Posts
Discussion Starter #12
The basic rules will be ohm's law.

So the actual voltage x the actual current will equal IMA assist power.
96V x 80A say for a standard setup would equal about 7.7kw max assist.
It might require the +40% current hack to get back to higher levels..
So lets say 96V x 113A (+40%) = 10.8kw

Regen might be more powerful as the motor voltage will be higher than the battery more of the time.

If it works at all we will see what's what.
 

·
Super Moderator
Joined
·
8,117 Posts
Driving the above might be achieved by an ICL8038 chip with op amp follower and a HV switching transistor.
I looked at the data sheet for the chip and checked some prices. Looks adequate and quite cost effective.
 

·
Administrator
Joined
·
11,556 Posts
Discussion Starter #14 (Edited)
I've been tweaking my bench G1 car mock up ready for testing.

Ordered one a sine wave generator chip to play with, might be a week before more progress.

I have salvaged 2x 400V 6.8uf caps from a dead psu for the voltage doubler..

Will build up the voltage doubler/BCM fooler combo and VPIN fooler in the meantime.

If the voltages synchronise and stay in step/proportion/etc on the bench we can test it in the car. :)
 

·
Super Moderator
Joined
·
8,117 Posts
The 10 x 10k 0.1% resistor BCM fooler subdivides that 150V into ten equal taps for our BCM.
I see no separate filter in the circuit. I assume that the effective RC time constant of the components is short and the primary circuit will provide ripple free operation???
 

·
Administrator
Joined
·
11,556 Posts
Discussion Starter #16
Ignore my earlier MCM HV Input schematic.

It has some serious errors like all the optocouplers the wrong way round!!

I'll post a new one shortly.

I've been doing a lot of looking at this today with a microscope,probes and oscilloscope.

I'm nearly at the point of it simulating in Proteus.. :)
 

·
Premium Member
Joined
·
806 Posts
We have had successful LTO 60, 72, 84 cell conversions, and even my proposed 96 cells as feasible.
retepsnikrep, I know we discussed 96 cells but how would you deal with the DC-DC converter which keeps powering off with high voltage?

Right now my 84 cells when charged to 2.46 volts will shut the DC-DC converter off when I brake at 2.46 volts. It is especially bad in the winter when the regen spikes the cells to 2.58 volts.

How could one use 96 cells and not deal with the constant issue of the DC-DC converter?
 

·
Administrator
Joined
·
11,556 Posts
Discussion Starter #18 (Edited)
Your question is not really for this thread.
This is about the 48 cell option and those specific technical hurdles.

However my answer is remove the OEM DC-DC and use a meanwell PSU.
Smaller, lighter and happy at the higher voltage as well.
 

·
Administrator
Joined
·
11,556 Posts
Discussion Starter #19
OK Hours spent on this (two days solid work) Reverse engineering the MCM HV input.

It now simulates using Proteus and the schematic has been simplified slightly by removing one optocoupler (which basically turn the HV detection on when the ignition is turned on) and a couple of resistors and an HV filter cap.

The components have been labelled using the MCM pcb ident numbers.
I have replaced the fixed resistors 237, 238, 241, 243 with variable types so I can experiment with values in the simulation.

Basically the 2 independent pwm signal drive 2 pairs of optocouplers to charge a HV capacitor and then transfer this charge to another capacitor on the LV side for measurement at approx 15ms intervals.

The output goes to an opamp which is probably conditioning the signal so it can be measured eventually by an adc of some sort.

In image 1 you can see a 150v battery gives a 1.30v peak in the pink trace.
84619


In image 2 you can see a 100v battery gives a 900mv peak in the pink trace.
84620


In image 3 you can see a 100v battery gives a 1.30v peak in the pink trace if we change RV241 to 316K :)
84621


We can add a resistor in parallel to 510K RV241 on the pcb to achieve the value required for our 50% voltage boost :)

Adding an 820k in parallel to our 510k will get us into the ballpark 316k approx.

Practical Testing now needed but it looks promising.
 
1 - 20 of 202 Posts
Top