BEWARE! I tried of of those fake Chinese AEM 30-0300 knock-offs, so you don't have to

Recently, someone sent me a "30-0300" to have its firmware re-flashed, for OBDII ( 30-0334 ) functionality.

It turns out, this is NOT an AEM 30-0300 wideband gauge, at all. It is a poorly made knock-off; probably from China. I looked around, and, saw that they are all over Amazon and eBay, as well as AliExpress.

Attachment 144684 Attachment 144686 Attachment 144687
FAKE on the left, real AEM on the right.
This fake has red buttons. Most of the fakes have black buttons. And, other than the missing ( usually ) AEM logo, they are physically indistinguishable from the real deal. Until you turn them on...

Attachment 144685 Attachment 144688
They copied the plastic parts, exactly. Except for the diffuser. All of the parts interchange, between real and fake.

But, the electronics are another matter. This thing is made SOOPER cheap. And, appears to have been designed in a hack-shop, that only needed it to "look like it works". And, it barely manages that much. Note the relative sparsity of components on the fake PCBs.

Attachment 144689
Ignore the dust. I had it out in the truck for road testing. But, note the display LED appearance, without a proper diffuser.
Also, note: There is no sensor connected to either gauge. But, the fake one goes through a fake "animation" of sensor detection, sensor typing, sensor heating, and, displaying AFR, anyway.

https://www.youtube.com/shorts/BADpXewRK9k

The display accuracy, when you can read it, is among the worst that I have ever tested. And, I have tested, pretty much, ALL OF THEM. Here are the results from a few calibrated lab test gasses:

GAS Actual O2% Actual λ Actual AFR Correct V Tested O2% Tested λ Tested AFR Tested V λ Error λ Error % AFR Error V Error
Free air 20.80% - - - - - - - - 4.500v 5.28% - - - - - - - - 4.875v - - - - - - - - - - - - 0.375v
8% O2 8% 1.241λ 18.18AFR 4.500v 5.28% - - - - - - - - 4.871v - - - - - - - - - - - - 0.371v
5% O2 5% 1.142λ 16.73AFR 3.965v 5.28% 1.040λ 15.24AFR 3.560v -0.102λ -8.93% -1.49AFR -0.405v
4% O2 4% 1.111λ 16.28AFR 3.774v 4.45% 1.250λ 18.31AFR 4.760v 0.139λ 12.51% 2.04AFR 0.986v
2% O2 2% 1.053λ 15.43AFR 3.416v 2.19% 1.125λ 16.48AFR 3.960v 0.072λ 6.84% 1.05AFR 0.544v
100% CO2 0% 1.003λ 14.69AFR 3.108v 0.55% 1.035λ 15.16AFR 3.375v 0.032λ 3.19% 0.47AFR 0.267v
BAR97H 0% 0.893λ 13.08AFR 2.429v 5.28% 0.945λ 13.84AFR 2.710v 0.052λ 5.82% 0.76AFR 0.281v

Note: 5.278% is the max clipping reading

All tested values are eyeball averages, as the readings are very noisy. AEM values are within 2% of actual.

The 0-5v analog output is nearly useless, even with the poor accuracy. The noise level is about the worst, that I have seen. It takes over 200ms to settle on a value, when the gas is switched. AEM, generally, less than 10ms.

Attachment 144690 Attachment 144698
It only updates 10 times per second, vs. 500, for the AEM. The 2-lane effect is from improper handling of the heater current; and, the lack of ground offset compensation. i.e. this thing shares the ANALOG OUT- line with the power ground! The rest of the noise appears to be digital switching noise from using a single power supply for analog and digital functions, and, the LEDs. Plus, it appears to use a PWM DAC, as the CPU that they use has no proper DAC, of its own. The actual voltages are a bit worse, even, than the display values.

The serial output seems to work OK. At least, it matches the display values. It loses a few messages due to noise. But, mostly, OK.

The CAN output is a mystery. It DOES send the CAN messages. But, there is no lambda/o2% data in them. Seems odd, to do all of the work to get the CAN messaging up and running. But, then, forget to encode the lambda data in the packets ?

There is a bunch of other technical crap. Like the complete lack of heater control, the harness wires are of too thin a gauge. The single power regulator runs EXTREMELY hot ( like 150?F in a 72?F room) . The gauge "locks up" every time you change a setting, requiring a power cycle, to recover. etc.

Using one of these to tune an engine is NOT a good idea. It gives enough of an appearance of working that you might trust it. Then, BOOM.