uM-FPUv3.1

caskaz · 2012-10-05 05:19

Hi.
I bought uM-FPUv3.1.
http://www.parallax.com/Store/Components/IntegratedCircuits/TimeMathCoprocessors/tabid/617/CategoryID/82/List/0/SortField/0/Level/a/ProductID/401/Default.aspx

I checked by objects below;
http://obex.parallax.com/objects/286/
http://obex.parallax.com/objects/813/
http://obex.parallax.com/objects/829/

But result are same;

---uM-FPU-V3.1 with 2-wire SPI connection---

FPU Software Reset done ...

Response to _SYNC: $FF   Not OK!
Please check hardware and restart...

I checkes I2C, too.
I changed SDA and SCL to P29 and P28.
http://obex.parallax.com/objects/287/

But fault

uM-FPU V3.1 and DS1621 demo with I2C protocol
I2C lines are High and STABLE!
---------------------I2C Scan-----------------
Scan Addr. : 0 NAK
....
Scan Addr. : 160 ACKNOWLEDGED <-- eeprom

...

No. of responding I2C devices: 1

DS1621 Missing!
uM-FPU Missing!

Trying FPU MASTER CLEAR...
uM-FPU Present!
FPU MASTER CLEAR...

-------------uM-FPU-V3.1 demo---------------
FPU Software Reset..
Response to _SYNC: 255  Not OK!
No Sync, trying a MASTERCLEAR first...
Then trying an FPU Software Reset...

I checked hardware many times and many times.
But no problem.
This chip is broken, or my mistakes?

Please any idea?

caskaz · 2012-10-05 06:35

Sorry, my mistake.

I connected uM-FPUv3.1's power to 3.3V on QuickStart board, because power is 5V,3.3V,2.7V on its datasheet.
But it seems uM-FPUv3.1 operate only 5V.

Duane C. Johnson · 2012-10-05 07:18

Hi caskaz;

Yes, the uMFPU-V3.1 is designed for 5V operation.
The I2C is also designed for 5V levels.
I suggest that you use an I2C bridging circuit as a go-between the uMFPU and the 3.3V og the prop.
Take a look at this circuit here:
http://forums.parallax.com/showthread.php?134122-Low-Cost-PropStick-variation-for-use-with-a-PlugBoard

Duane J

Publison · 2012-10-05 08:36

Interesting. It seems I.Kovesdi got it to work at 3.3 V with his I2C Driver Demo. I don't have a DS1621 right now, or else I would try it.

Duane C. Johnson · 2012-10-05 10:54

Well I was looking for reasons why it wasn't working?

The spec for the uMFPU-V3.1 calls for a minimum logic high of 4V with a VDD of 5V.

Of course many chips will run properly at less than this. I suspect yours does, but without the I2C bridge it is still outside of the spec requirements.

Duane J

Duane Degn · 2012-10-05 10:57

Duane C. Johnson wrote: »

I suggest that you use an I2C bridging circuit as a go-between the uMFPU and the 3.3V og the prop.

Duane,

I don't see why any level shifter is needed. As long as the I2C object you're using is obeying the I2C rules of only pulling the lines low it shouldn't be a problem using a 5V I2C device with the Prop. (I think there is a Propeller board that doesn't use pull-ups on both I2C lines but IIRC, there is only one board (sold by Parallax) that doesn't obey the I2C rule of pull-up resistors on both the data and clock lines.)

The I2C lines could be pulled high (to 5V) wth a 10K resistor which would keep the current low enough not to cause a problem with the Prop. Many 5V I2C devices will work with the lines pulled up to 3.3V as well. I personally think since 5V (on the I2C bus) through a 10K resistor doesn't present a problem for the Prop, one should use 5V and then they would be sure it would work with any 5V I2C device.

I could see how voltage differencese could be problem if you had 3.3V I2C devices that couldn't tolerate 5V even with very limited current and 5V devices that don't recognize 3.3V as logic high on the same I2C bus. But I doubt this senario occurs very often (at least at the hobby level).

I know I've used the uM-FPUv3.1 (and other 5V I2C devices) with the Prop several times without any sort of level converter.

Duane Degn · 2012-10-05 11:21

Duane C. Johnson wrote: »

but without the I2C bridge it is still outside of the spec requirements.

Duane,

It makes me nervous not to be agreeing with you since I know you know a lot more about this stuff than I do. So I'm hoping you'll correct me if I'm wrong about it being safe to use a 5V I2C bus with the Prop. I'd think as long as the pull-up resistors keep the current below the 500uA limit the Prop's clamping diodes have, it should be safe to with the Prop?

Duane C. Johnson · 2012-10-05 15:06

Correct, I was just commenting on the communication aspects of his possible problem.

A 3.4K resister, commonly 5K, would limit the current to 500uA for safety assuming 5V and 3.3V.
5K is safer in case the voltage is 5.5V and the prop is running at 3V.

Yes, many have found the uMFPU-V3.1 works with the simple resister pair pullups to 5V. Mine didn't though. I needed the I2C bridge.

Technically the simple resister pair pullups don't pullup to 5V. The voltage only gets to about 3.3V + .6V = 3.9V or so. Pretty close to the required 4V.

But seriously, the simple 2 MOSFET circuit goes to very near 5V and would allow lower valued resistors for faster rise times.

Duane J

Duane Degn · 2012-10-05 15:37

Duane C. Johnson wrote: »

Correct, I was just commenting on the communication aspects of his possible problem.

A 3.4K resister, commonly 5K, would limit the current to 500uA for safety assuming 5V and 3.3V.
5K is safer in case the voltage is 5.5V and the prop is running at 3V.

Yes, many have found the uMFPU-V3.1 works with the simple resister pair pullups to 5V. Mine didn't though. I needed the I2C bridge.

Technically the simple resister pair pullups don't pullup to 5V. The voltage only gets to about 3.3V + .6V = 3.9V or so. Pretty close to the required 4V.

But seriously, the simple 2 MOSFET circuit goes to very near 5V and would allow lower valued resistors for faster rise times.

Duane J

Thank you for taking the time to explain this. I appreciate it very much.

nglordi · 2012-10-05 16:51

caskaz:

The uM-FPU3.1 does work with VDD set to 3.3V. I would suggest that you try the ProForth SPI FPU word set posted on the PropForth wiki. The 4V minimum for a high only applies if VDD is 5V.

NickL

caskaz · 2012-10-05 20:22

Hi.
It seems uM-FPUv3.1 works on 3.3v power.
Chip returned reply($5C) by sending SYNC($F0) after reset.

Firstly, I entered to debug monitor.
I changed CLOCK to 14.74MHz.

>$C
E3
:EA

uM-FPUv3.1 works 29.48MHz under default.
When 3.3V power , chip fail back to internal clock speed of 1.8425MHz because of no operating on 29.48MHz.
Maybe speed is slower than 5V power.

This question is FAQ?
Everybody know?
I had never known this.

camt · 2012-10-06 02:55

Hi,

The maximum clock frequency of the uM-FPU V3.1 chip is limited by the operating voltage. The factory setting is 29.48 MHz for operation at 5V. If you want to use the chip at 3.3V, you need to change the clock speed to 14.74 MHz. You can do that with the uM-FPU V3 IDE software or manually using the debug monitor (see the datasheet $C command). You should always power cycle the chip after changing the clock speed, then check the version string to confirm that it is operating at the correct clock speed. The 1.8425 MHz clock speed is a default clock, which allows the debug monitor to run if the specified clock setting fails. I would recommend using a 5V supply while you make the change, then use the 3.3V supply after the clock has been set properly.

SPI and I2C will both work at 3.3V.

Regards,
Cam Thompson
Micromega Corporation

uM-FPUv3.1

Comments