rusefi.com

We need some H-Bridge so why not TLE7209 or MC33186, these seem to be compatible enough and some older versions available from China?

TLE7209-3R has recently replaced TLE7209-2R seems to be pin compatible.

See https://github.com/rusefi/hw_modular/tree/master/TLE7209_or_MC33186_H-Bridge_Breakout

iBom https://rusefi.com/docs/ibom/TLE7209_Module_latest.html

Why would we ever use an NRND part when there are better parts available for less money that are still active? Seems pretty foolish to pick a random part without any research, especially one that’s NRND.

For example the ST L9960 is designed for ETB applications, automotive rated, fully protected, adjustable over current protection, SPI fault detection, etc. Oh, and they’re <$5 in single quantities. And there’s also a variant that’s two chips in the same package to save space if you need to drive two motors.

mck1117 wrote: ↑

Mon Mar 11, 2019 5:25 am

Why would we ever use an NRND part when there are better parts available for less money that are still active? Seems pretty foolish to pick a random part without any research, especially one that’s NRND.

Top message corrected - there is a TLE7209-3R not NRND recent revision.

PS: I am always open to any other suggestions. As someone begging for any help, KiCad help included, I would love to have any other H-Brigde breakout board, ideally with a two-row side connector.

mck1117 wrote: ↑

Mon Mar 11, 2019 5:25 am

Why would we ever use an NRND part when there are better parts available for less money that are still active? Seems pretty foolish to pick a random part without any research, especially one that’s NRND.

For example the ST L9960 is designed for ETB applications, automotive rated, fully protected, adjustable over current protection, SPI fault detection, etc. Oh, and they’re <$5 in single quantities. And there’s also a variant that’s two chips in the same package to save space if you need to drive two motors.

I dont think that TLE7209 is picked random without any research.

That's actually the driver found in each and every Bosch-ECU for almost 20 years now.

Ah okay. The OP made it sound like it was already decided, and provided no context or reasoning

The Infineon and ST parts look quite similar functionality-wise, but the Infineon is kinda hard to source. I'm not a fan of sketchy ebay parts. The older 7209-2 is available from normal places, but it's NRND. The new -3 isn't in stock anywhere (yet?).

The interface on the L9960 and TLE looks similar enough that they may be compatible (save for SPI commands and stuff), so I tacked a few L9960 on an order I was already placing.

@russian : what kind of connector is useful on the edge of the board for you?

mck1117 wrote: ↑

Wed Mar 13, 2019 12:59 am

@russian : what kind of connector is useful on the edge of the board for you?

For this to be used on the blue nucleo wing it would be great to follow the https://rusefi.com/forum/viewtopic.php?f=4&t=450&p=32039#p32039 and https://rusefi.com/forum/viewtopic.php?f=4&t=1471#p31990 using https://rusefi.com/forum/viewtopic.php?f=4&t=1473#p32038 header

mck1117 wrote: ↑

Wed Mar 13, 2019 12:59 am

Ah okay. The OP made it sound like it was already decided, and provided no context or reasoning

The Infineon and ST parts look quite similar functionality-wise, but the Infineon is kinda hard to source. I'm not a fan of sketchy ebay parts. The older 7209-2 is available from normal places, but it's NRND. The new -3 isn't in stock anywhere (yet?).

The interface on the L9960 and TLE looks similar enough that they may be compatible (save for SPI commands and stuff), so I tacked a few L9960 on an order I was already placing.

@russian : what kind of connector is useful on the edge of the board for you?

The7209-2R are still in stock at both Mouser and Infineon.

The 7209-3R looks to be just a new revision, and it might be they want to sell out the 2R before releasing.

NXP also have newer versions than the 33186 that are identical.

These seems to have become the "standard".

So by using them, we are sure they works perfect with all E-throttles made the last 15-20 years.

(Standalone ECU's like LINK and MaxxECU use it as well)

Just ordered first batch at jlcpcb

Hi there,

a bit late to the show, did not have rusefi on my radar sadly.
The TLE is the chip, as i am mainly involved in reversengineering Motronics for a long time i have not seen another ETB driver used in all those years.
Early Motronics (ME7.1,ME7.1.1,ME7.3.x) use the SF flag feature, newer ones use the SPI.

Just wanted to add that you should add a largeish cap near it as otherwise it might become quite erratic.
Most of them use a 220uF electrolytic cap near its supply, only the Hybrids use a 10uF.

Hi, bit late to the show.

But you made the right choice, i have never seen any ETB motronic without that chip.
What they tend to do is what you also should is adding a large CAP to the supply of the chip.
They mostly use a 220uF cap, only the hybrid versions have a 10uF cap, and they are often failing.
Looks like surges can somehow kill that chip.

In older motronics they use the SF output, newer ones use the SPI, both is fine imho.

We have recently damaged too many of these 7209. A couple from ebay and a couple from mouser.
Maybe this cap hint is what is killing these chips?!

Thank you @960 for adding large capacitor - I'v just ordered a batch of pre-assembled PCBs.

I uses a pololu board for my throttle and they say add a large cap...I didn't and the board died. I bought a new one with higher current rating and add the cap and it seems good now so one or the other fixed it....but they say use a large cap so I'm guessing that's the answer.

So here's why the cap is required. Without a cap, the circuit forms an LC resonator (essentially a boost converter!) that can generate extremely high voltages at the tle7209's power supply.

Here's a spice model, and corresponding simulation:



Blue is the supply voltage, at the modeled tle7209. When the load is turned off, inductor L1 (the supply wires - not a separate inductor) tries to keep flowing current. C1 absorbs some of it, but putting the energy from that inductor in to the cap results in a really high voltage.

Here's the same test, with the addition of a 220uF, 0.1ohm ESR electrolytic capacitor:



Now there's less than a volt of ripple, and no massive spike (nor is there a huge drop when the load turns on).

This is an approximate model since the losses are difficult to model, but it gets the general idea.

I would recommend using a ceramic capacitor. I typically have not heard or seen electrolytic's used as bypass caps for ICs since they have higher inductance (ESL) which is what you are trying to get rid of with a bypass cap. If you wanted the best you could go with a tantalum cap but why spend the extra money if its not needed?

Also the application circuit for the TLE7209 shows 2 bypass capacitors, a 100 nF and a 100 uF.

SVeilleux9 wrote: ↑

Sun Jan 12, 2020 1:46 am

I would recommend using a ceramic capacitor.

That would be great, except that 50v 220uF ceramic capacitors don't exist. Essentially all OEM ECUs use electrolytic bulk caps for both the main supply and main relay "power device" supply (ETB, stepper idle, GDI boost supply, solenoids, etc).

SVeilleux9 wrote: ↑

Sun Jan 12, 2020 1:46 am

I typically have not heard or seen electrolytic's used as bypass caps for ICs since they have higher inductance (ESL) which is what you are trying to get rid of with a bypass cap.

That's correct, there's no reason to use an electrolytic as a bypass cap, since they're never more than a few uF at a few volts, and ceramics have lower series inductance and resistance (ESL/ESR). However, 100uF is not a bypass capacitor, it's a bulk capacitor. That said there's really no quantitative difference.

Here's one way to think of it: The job of bulk/bypass capacitors is to provide a local supply with lower resistance and/or inductance than the next supply up the stream. The wires from the fusebox/main relay to the ECU have a loop resistance of perhaps 25 milliohms, but a huge inductance of up to a few microhenries (model above is 10mOhm and 1uH). A typical 100uF aluminum electrolytic will have a typical resistance of perhaps 100 milliohms, but a relatively tiny inductance of only tens of nanohenries (2 orders of magnitude smaller than the wire).

If we really really cared about extremely stable voltage, maybe it would be warranted to stack tens of ceramic caps, but we sort of don't. The 100-150mOhm ESR of the electrolytic is absolutely acceptable for the cost and size savings it offers.

Revision 0.1 board could be saved for prototyping purposes by hard-wiring C4 with wires below 2cm, and R4/R5 values reduced. We will soon generate 0.3 PDF with all corrections.