Archives

I would not be the first to recognize that some things in life are just bound to take a back seat to everything else. Some of you may know what I am talking about: Hobbies, vacation, that long rest you promised yourself after you complete "the big project"?

Well, it looks like in my case I will not be able to have my IMA assist inhibit module 2 complete before the end of the year. Too many things are happening on the business side of life are going to consume much of my attention and time.
So, instead of procrastinating indefinitely, I figured I would open an invitation to all those interested to review (and discuss) a far more simplistic approach to implementing this mod on an HCH-2.

The Objective:
In this invitation I will offer details that will clearly identify the means through which to cause an assist inhibit condition.

Why an inhibit condition? What is the benefit?

As some of you already pointed out, it is very difficult to prevent EV assist from taking place under normal operating conditions on a stock HCH-II. Our high mileage performance strategy may consist of saving the SoC for times when it can yield the greatest benefits. For instance, in the summer I will often prevent assist so that when I operate the AC, I may suffer a lesser FE hit.
Another condition may occur when we engage in higher performance P&G. In this situation it is nice not to use any SoC during the pulse, and this way the frequency of forced regens is greatly minimized.

Anyhow, as some of you may recall we've managed to produce a couple of very complex assist inhibits. My 06 HCH was the lucky recipient on one of those modules and while it works OK, it was also extremely expensive. To make matters worse, the final cost of the unit could not be offset from the benefits that it offers mostly because I already have a natural assist inhibit: the winter weather (or more precisely the low temps)!
During the colder winter months, we often experience minimal EV assist so an assist inhibit mod is literally a wasted feature. Well, I've been thinking that since the weather can play it that way, perhaps we can use that approach as the basis for a simpler and less intrusive "assist inhibit module". Get my drift?

Furthermore, I've had the opportunity to communicate my thoughts to a small group of folks at CleanMPG over the last 6-7 months and in those instances I did reveal that there may be a couple of ways to implement a much easier mechanism to a microprocessor based mod.

The approach:
So here it is: A descriptive blueprint to a simple and inexpensive assist inhibit for the HCH-2. I may produce and publish a multisim schematic for those who request it.


What we will need:
-Experience and comfort around electronics and high power electronic equipment.
-Precision Multimeter
-Insulated rubber gloves
-basic tools, pliers, cutters, crimper and connectors, soldering gun, heat shrink tubing, etc
-3 relays (NC) + 3 attenuation diodes. You may also use solid state relays if you wish. :)
-NO push switch to install inside the dead pedal
-3 high precision resistors (will discuss the value and precision later) ;)

So what will this mod consist of?
We must power OFF the IPU, open the cover and locate the BCM connector module 40P.

This connector contains 3 pins that are of interest to us. These pins are:
pin 33 - TBATT1
pin 34 - TBATT2
pin 15 - TBATT3

You guessed it. These are the pins that provide a voltage reading of 0.05V (90C) to 4.95V (-40C).

The plan:
Now the principle here is that if we throw the readings to... let's say a VERY -ve range then the BCM will inevitably tell to the MCM that there'll be NO ASSIST. So, the key here is to manipulate the voltage reading on all three pins simultaneously and by precisely the same amount.
Some will say... hummm... why not just set the reading of a single pin instead of all three at the same time?
Well if we do it that way, then this will generate a code since the BCM is sees that as a temperature sensor malfunction and we definitely do not want that condition to occur.

Anyhow, the reading on these pins can be affected in two easy ways:
- Introduce a resistor (value TBD) to each of the pins and connect it to pin 6 on the connector (this is a BCM ground). This resistor will be one of three that must be activated by each relay simultaneously.
- Introduce a resistor in series (value TBD) to the pin by cutting the wire pins before reaching the connector so that resistor may work as a "value shifter". Again, this resistor will be one of three that must be activated by each relay simultaneouisly.

Please note the voltage values I noted. We must remain within 0.05 and 4.95 Volts !!!

Anyway, I like the first option because I know it works from early May 2006 testing. ;)

--------------------------------------

So how would it all work?
Under normal conditions the circuit will work as stock. However, when we press the dead pedal button the normally closed relays will be triggered and the resistor values will be introduced, thus shifting the temp readings of the battery. This in turn will cause the MCM to deny assist as a denial condition was just created.

Concerns:
The BCM also monitors and drives the fan control unit which regulates the temperature of the IPU. From experience, we know that this too is taken as an input that can cause an EV exception/shutdown condition.
In order for us to avoid this, it is necessary to limit the amount of duration of these assist inhibits. In the hands of a knowledgeable driver this may not be a problem, but if a kit was ever produced that could be installed by any end user, then it would be necessary to include a driver circuit that would effectively "time" and cut-off the inhibit requests. From my experiments, I've found that inhibits lasting less than 20 seconds are OK.

Conclusion:
I am hoping that this theoretical incursion will provoke those daring souls (you know who you are - especially those with whom I've been talking over the last 10 months;) ) to help me validate this approach in the hope that we can turn this into a collective project by CleanMPG members.
Because I no longer have the time or opportunity to continue working on this, I figured this would be the best way to research and validate an easier mod that could be made available to any HCH-2 owner seeking an extra MPG edge.

I have more details and information concerning parts and additional support circuitry, but I will withhold that information until a little later. This is because I am still trying to line-up a few days off for "playtime" sometime in August. :)

In the meantime, I encourage anyone interested to apply any scrutiny and ask any questions about what I discussed above. If you wish, you may also PM me and I will reply in kind.

Cheers;

MSantos

Hi MSantos:

___I have been waiting for a very long time to see an Assist inhibit switch on an HCH-II and in particular, your own design. I know you had a ton of safeties installed on your premium circuit but this one sounds a lot simpler? It makes P&G practical for one and two, does everything you have already stated above.

___Although I do not own an HCH-II, would breaking the circuit to those 3 pins (yielding 0.00 V) via a timed Relay drive the Assist inhibit signal via fake temperature input for the 20 seconds or does it need to see 0.05 V as an absolute minimum on the three pins at the same time? Using a relay with an internal time delay before reset takes away the High temp fan triggers from the equation? 20 seconds is all I could ever imagine needing for a pulse myself so it would work out really well!

___Good Luck

___Wayne

As always Wayne, you are "dead on" !!

Yes, the system needs to see the 0.05 V, otherwise it is likely interpret the "state" as a failure (quite interesting to hypothesize nevertheless). In all honesty, we have not really determined what the hot threshold is. For that, we would have to either experiment a little more ;) or look into the BCM software. The only other way to "kind of see" the threshold would be to plug a CAN-VIEW like device (like the HDS station) and look into the CAN stream. But then, we would have to create the conditions for the experiment and removing a temp sensor for a quick torch bath is not that easy because it would require deep tampering into the battery. In fact, Honda simply says that if a sensor is malfunctioning then the whole battery pack must be replaced, simply because they do not expect/recommend dis-assembly of the cells?!

Anyway, depending on the switching speed of the relays (whether solid state or not) there will be "technically" a fast transition of values on the voltages for each of the 3 temp sensors wires. They key is to keep the transition as small and smooth as possible.

Here's the other caveat: Software updates.

I've had a few on my 06 for just about every module since I purchased the car in May/June 2006. So keeping an eye on the potential software differences (however small) between HCH-II's is very important; that is why having a larger community of owners participating in such an endeavor is a much safer approach in addition to speeding up the dev process. Yes, I do have 2 HCH's but my 06 is now at the same software baseline as the 07 and the thing about this Honda software is that you cannot rollback to an older version. Heck, you can't even reload the same version. Any software update will only occur if and only if a new version is found or a new/replacement module is introduced into the IPU.

Cheers;

MSantos

Is it 0.05V or 0.5V? Usually sensors like that sit at or
near one diode drop from ground or + as a valid "control
range", and wandering outside that indicates a short.
.
_H*

Hobbit:

It is definitely 0.05 volts. The sensor is of resistive type and it connects the BCM ground to the BCM input.
Indeed, wandering outside the voltage ranges I mentioned, could indicate a short or an open on the sensor. In any case, EV operations cease well before reaching either of those values anyway.


Cheers;

MSantos

Somebody PM'd me the other day wondering about the prospects of a MIMA like device for the HCH-II. :rolleyes:

Perhaps Mike can help me understand what MIMA does for our Insight owners, in case I have it wrong: MIMA allows manual control of the Regen and assist functions correct?

If the above is roughly true then how difficult would it be to come up with something like that for the HCH-II?

This is what I know we can do without Reading and Writing to the CAN BUS:
1 - We can prevent assist :)
2 - We can force a regen :)
3 - We cannot easily force assist :(

Explanation:

1 - Preventing assist could be done quite easily. Please see my previous posts.
2 - Forcing a regen is also quite "easy" too. I'll explain this one below.
3 - An assist condition can only be created if we "write" to the CAN BUS. As soon as we consider this approach, we are implicitly looking at dedicated digital circuitry as well as a requirement for the protocol specifics, state values and ranges, device descriptors and bus device IDs and so on.
Additionally, unless someone gets the necessary "Gen 4 IMA inside information" (yeah sure :rolleyes:), it would also be necessary for us to record a full CAN stream and then analyze it in an attempt to decode and interpret the most pertinent communication details. Quite a lot of fun this reverse engineering stuff, heh?


OK, what about number 2 (forced regen) then. How would that work?

Simple. The BCM connector has the infamous 40P connector which we already focused on previously, and this time we'll identify the following pins that will provide the implementation means:

PIN 16 - SOC + value which ranges between 0.3V and 5.0 V
PIN 36 - SOC - value

Now, since these voltage values are derived though induction via the battery current sensor(in the junction board) the resulting voltage levels indicate a low SoC at 0.3V and a high SoC at 5.0V.
To force a regen condition we would have to alter the voltage values as they emerge from the junction board to Pin 16 on the 40P.connector.
Although this can be done in any number of simple ways, I would only dare to do so with safety in mind. In other words, providing too much regen could lead to an overcharge condition and that is a no-no in my books. I would only be comfortable with the idea if the manual regen requests could be inhibited in case these lead to an SoC of 80% or more. In order to make this happen we would need monitoring circuitry that would either allow or deny manual regen requests as long as it can operate within a good margin of safety.

Any additional thoughts on a Forced Regen mod? Please feel free to contribute. ;)



Anyway, what does this all mean?

Assist inhibit functions are the lowest hanging fruit, as the mod can be cheap and relatively simple.
Forced Regen functions are possible but will likely require extra design/engineering effort for safety reasons.And a few "take away" questions:
Could a potential mod include both features?
Should it just contain the Assist inhibit?
What is the point of inducing a forced regen? I can think of a couple of good reasons but I doubt most forks would regard it as useful enough to justify the additional engineering and material costs. Thoughts ?Cheers;

MSantos

MSantos--

I'm no engineer, but I think I've kept up so far with your discourse and I'd like to offer input if I can.

I think an Assist Inhibit function would be the most useful of all the mods, as you've stated. HCH drivers have a couple of tools at their disposal to invoke regen, whether it's via brake pressure modulation, "S" mode on the CVT, etc. While these aren't perfect analogues to your scheme, they present at least an alternative.

One of the reasons I haven't employed P&G to much success is that I have yet to master to Assist-less pulse. Therefore when I try it I get hit with many forced recharges. It's really not easy to accelerate without assist at speeds below 35 and with any amount of hill in front of you.

Anyway, I'm rambling -- Good luck and let me know if a layman can help. --RN

Hi MSantos:

___One of the ways we can force IMA equipped hybrids (the Insight, HCH-I and II all act the same) is to bring their SoC down to 1/3 of the display and they will stop assisting on every pulse. The caveat is they will force charge on every pulse too. We accept that as given because it actually helps maintain a minimum level of SoC to cover the overhead. The problem of course is when your P&G cycles are complete, they will force charge back up to 19 of 20 or 7 of 8 and that payback is a killer! It is one thing to P&G for 30 miles or so but very few can do that let alone the horrendous payback to bring her SoC back :(

___If you could attack the Assist inhibit from that angle without the forced charge to keep the pack from seeing its maximum SoC limits, that could work too?

___I like your first mod with the time inhibit but the above is another way of killing the bear with the negative that the hunter has to shoot himself in the foot after the hunt is over :(

___I get mildly perturbed just thinking about the Honda engineers that designed the “Assist on every accel for performance” scheme. Those micro-hits on the NiMH have to be horrendous for an all-city driver in particular. IIRC, the Prius-II pack was meant to be touched (Regen and Assist) around 250 - 300,000 times with these mini-bursts but they are easily limited and controlled by both Toyota and the driver. The Honda’s with their far less robust NiMH’s must be taking a beating with the ultra-sensitive Assist.

___Good Luck

___Wayne

Wayne;

I agree that the HCH-II is not in its best element at all when driven exclusively in city streets. In the hands of a regular driver it will really take a good FE beating.
In contrast to that, the Prius is far more resilient and thus born for that type of environment; in the hands of a regular owner it will always achieve good mileage.

No contest so far.

After a year of living with the HCH-II, and driving it in city-only routine, I've learned a few good things and a few bad things about it. The absolute worst bad thing about the HCH-II is that the 1.3L I4 is a real dog when electric assist is absent. I mean, consider the combined power and torque of both power sources: while adequate for most folks, it cannot even come close to being adequate for some.
Now, take away the assist and the car accelerates like a whining snail. So, to some extent I can understand why Honda engineers made the assist so sensitive and ever present. I believe they really expected to target a more suburban/highway driving profile where the car would excel in steady state driving.

Of course, this assist sensitivity can also be a hassle and removing it out of the equation for those unique circumstances is absolutely necessary. That has been my realization since I bought my 06.

http://www.cleanmpg.com/photos/data/522/power-torque_curves.jpg


On the subject of the batteries:

We've had a chance to compare the NiMH packs from both the Prius and the HCH-II (on the bench and on the available schematics front) and other than the size/capacity/package-design, we did not see any significant operational engineering differences between the two. So, maybe I am missing an attribute or two that would make the HCH-II NiMH less robust. I can't speak for the NiMH on the insight or HCH-I. Is it the cell shunting feature you're talking about ?


Cheers;

MSantos