DIYEFI.org Forum

lol, that is exactly what I do on SLC, the AC signal period is twice that of the ADC integral time, the difference (delta) between the samples is the nernst resistance the mean of both samples is the nernst voltage. Delta feeds into the PID heater control loop, mean feeds into the pump cell PID.

One the scope the symmetry of the noise does not look totally symmetric when you compare the volts vs time, due to the difference in the FET rise and fall times, but if you look at it noise power it is pretty symmetric. You can force things to look more symmetric aesthetically on a scope by rigging up a diode circuit on the gate so that the rise and fall times are roughly the same. I have the foot prints for that on the SLC PP2 design, but I never needed to play that card as it never added any measurably better noise performance with the circuit in place.

As long as the integral =0 it's all good

Here is the AEM inline UEGO PCB

CJ125 based. But seems to be a bit more going on than in the reference design.

i.e. The CPU is a PIC24FJ32 32MHz. That seems like a lot of CPU for a design that already has the CJxxx chip. I think they are doing something fancy for the heater control.

There are two separate voltage regulators ( hidden under the black gunk )
Both feed from the big diode and aluminum cap. Probably 5.0v for the CJ and 3.3v for the PIC.

The layout actually looks pretty good ( especially, compared to the Innovate and PLX parts )

If you look closely, it looks like, at some point, they changed the color coding of the external wires

Some of the hand-soldering of the wires looks a little ham-fisted. But, they left a lot of guard-gap clearance around the pads ( maybe for just such an occasion )

What about the ECM meters?

Like the AFM1600L...



I've read that AFM1600L and NGK AFX are the same...

Pardon the thread necromancy.

I think I found the missing link with the LC-2 and the switching regulator. J

I had an old dead XD-16 lying in the drawer. While moving stuff around, something caught my eye. The XD-16 had the same switching regulator as the LC-2. It even had the same related component values. The XD-16 was an all-digital display-only device. So, it had no need for a separate linear regulator for analog circuitry. It DID have a big bi-metal resettable fuse and a big TVS diode on the power input side.

It had 21 full RGB LEDs, so, it likely needed all the current it could get. Hence, the 1.5A switcher. All good.

This made me think of the MTX-L. The MTX-L was basically a combining of the LC-1 and XD-16 devices into a single unit. So… I opened up my MTX-L, to find :
It, again, had the exact same switching regulator, inductor, and diode as both the XD-16 and LC-2.

So, the whole switching regulator circuit was cut-and-pasted from the XD-16 to the MTX-L, which may have similar current requirements. But, still, that LT1767 is ancient.
The MTX-L does not seem to have a separate linear regulator for the analog circuitry. So, it has, not only, the switching regulator noise,; but, also, the LED multiplexing/switching noise, right on the 5v power rail of the analog circuitry.

The LC-2 came out about 3 years after the MTX-L. And it has the very same power supply circuit cut-and-pasted onto it. The LC-2 seems to be more closely related to the MTX-L than it is to the LC-1. i.e. it appears to be an MTX-L, less the display PCB. It has the same PWM DAC, no MAX202 for serial, etc.

I think Alan is right. There is probably les than 50ma being pulled through that switcher on the LC-2. It was just cut-and-pasted because it was on the previous design. This makes me think that whoever designed the MTX-L and the LC-2 has no idea of how it actually works.

At least, on the MTX-L, the switcher may be operating in the continuous range. On the LC-2, it is likely “hit and miss”.

Reading Innovate’s info on their controller circuit, it is pretty obvious that it is VERY sensitive to noise spikes. Dollars to donuts, this is a big part of why all of the newer Innovate widebands can not run in “instant mode” like the previous LM-1 and LC-1 which had cleaner linear power sections.

Note components A, B and C on each PCB
.

You would recommend the MTX-L or the LC-2?

Evolution-VII wrote:What about the ECM meters?

Like the AFM1600L...

I've read that AFM1600L and NGK AFX are the same...

the afm1600l and ngk afx share the same setup, difference is firmware.
Call ECM co and ask them, I have.

You would recommend the MTX-L or the LC-2?

Based on what I see here, I can not recommend either of them.
If you want an Innovate product, I suggest finding an LM-1 or older LC-1. Then, make damned sure that you keep the sensor from overheating. Maybe with one of the HBX-1 heatsinks; and positioning the sensor down-stream, as far as is practical. People who do this seem to suffer much less from the "Error-8" issue.

Maybe Alan can check me on this. But, I believe that Innovate runs the sensors colder than stock. So, sporadic higher EGTs , on an unprotected sensor, cause greater temperature swings than normal; and, possibly, sensor damage.

Otherwise, I would say, go with one of Alan's designs. He seems to have a handle on this stuff.

go with one of Alan's designs. He seems to have a handle on this stuff.

+1 :-) rocking one of his units on my car every day!

Also, thanks for more outstanding posts! Really appreciated :-)

Fred.

Hentai wrote:

Evolution-VII wrote:What about the ECM meters?

Like the AFM1600L...

I've read that AFM1600L and NGK AFX are the same...

the afm1600l and ngk afx share the same setup, difference is firmware.
Call ECM co and ask them, I have.

But the NGK do have the same quality though?