TPIC8101 Advance SPI Mode??

Ed,

In your first paragraph, it sounds like you are having trouble getting it to work at all, is this true even when using the normal mode for SPI communication?  What settings are you using?

Are you using your own board and microcontroller?

I'll take a look at the advanced SPI mode again tomorrow to see if I can offer you any advice that might help.

Thanks,

-Clancy

Clancy, thanks for the response.

To be honest, I am not certain. With the TPIC8101 in Normal SPI mode, the unit echos in the returned byte response the same command I send it. From my understanding, this indicates that my microcontroller (PIC32MX795F512H) SPI and the TPIC8101 are communicating in normal mode.

For example if I send the following to the TPIC8101
Prescaler: 80 I receive the value 80 back and display it on my PIC32 LCD display
sent received
Channel : 224 224 (channel 1)
BandPass: 30 30
Gain: 130 130
Ts: 222 222
Adv Mode: 113 113

My understanding is that the last command 113/0x71/1110 0001 should put the TPIC8101 into advance mode which is necessary in order to get the integrator output value via SPI. Well, as I mentioned, I am not getting any returned integrator output data.

So that naturally raises the question as to whether the TPIC8101 is truly in Advanced SPI mode. Then I found item No. 6 in table 2 (top of page 14 in my data sheet) which indicates that if 113 is sent to the TPIC8101 when it has already been put into Advanced SPI mode then instead of the echo of 113 I got in normal mode, I should get an "inverted SPi Configuration (MSB) 1000 1110 (LSB).

I send the 113 command after the board is powered up when the SPI mode is in Normal by default and I get the expected 113 response. However, when I send 113 again (when the TPIC8101 is supposedly in advance mode), I should get back 1000 1110/
142, but instead get back the 113 indicating the chip is not in advanced mode which of course means I won't get data over the spi link.

Yes, I do have my own board and microcontroller to answer your question. The board had been proven thought a number of units. The only thing different this time is I added the TPIC8101 and its components to the board.

I do get analog output whenever I take the INTHOLD pin high for approx. 3msec and then low. The analogy output shows a 7 step ramp - however, it shows this whether a "Knock signal" is present or not - which I understand from other posts is normal. I am using a function generator to generate the knock signal a sine wave 300mpVPP which I can then modulate with a square, triangular or sine wave patter or no modulation at all.

I have a OSC clock lead off my microcontroller resonator circuit which outputs a 200mv PP 24mhz signal to the TPIC8101. I have measure it at the pin to the chip and it is present. I have tried using both input channels and have measure the signal traces to make certain there is no broken connections.

A very neat chip which would enhance my product to my customers - if I could just figure out what I am doing incorrectly. I can see the SPI clk, SDI and SDO traces on my scope and all looks normal to me.

So if you can spare the time, I would greatly appreciate it.

Here is a segment of code that I use to set the values for the various parameters into variables which are then sent via SPI to the TPIC8101


//KS Utilities in SPI Advanced Mode - enter parameter value in decimal format
Procedure Set_Prescaler(Parameter:byte);
begin
//parameter entered is freq in mhz of Timing clock input to TPIC8101 (24mhz in this case)
//This assumes you want SDO of TPIC8101 active for SPI communciation
case Parameter of
4: PRESCALER_KS := %01000000 + %00000000; // 64 + 0
5: PRESCALER_KS := %01000000 + %00000010; // 64 + 2
6: PRESCALER_KS := %01000000 + %00000100; //64+4
8: PRESCALER_KS := %01000000 + %00000110; //64+6
10: PRESCALER_KS := %01000000 + %00001000; //64+8 = 72
12: PRESCALER_KS := %01000000 + %00001010; //64+10 = 74
16: PRESCALER_KS := %01000000 + %00001100; //64 +12 = 76
20: PRESCALER_KS := %01000000 + %00001110; //64 + 14 = 80
24: PRESCALER_KS := %01000000 + %00010000; //64 + 16;// PRESCALER := %01000000 + %00010000; //64 + 16 =80
end; // case

end;
Procedure Set_CHSEL(Parameter:byte);
begin
if parameter = 1 then
CHSEL := $E0; //Channel 1
if parameter = 2 then
CHSEL := $E1; //Channel 2
end;
Procedure Set_Bandpass(Parameter:byte);
begin
Bandpass := 0 + Parameter;
end;
Procedure Set_Gain(Parameter:byte);
begin
Gain := 128 + Parameter;
end;
Procedure Set_Ts(Parameter:byte);
begin
Ts := 192 + Parameter;
end;

Procedure SPI_Advance_Mode(Parameter:byte)
begin
Advanced_mode := 113;
end;

These values are sent as initialization and then an
Display_Name := SPI_Read(SPIBUF); is used to read the PIC32 SPI buffer with the echoed value in it and display the response. A different display name for each variable the value of which is read immediately after the command is sent.

The logic of the routine goes something like the following:

SPI_Write(BandPass); //send command
BandPass_Display := SPI_Read(SPIBUF); //read and store response value
Display(BandPass_Display); which later displays the results

I feel like I am close, but clearly missing something crucial

Sincerely

Ed

Hi Clancy,

Based on a recommendation made by John Griffin back in 2011, it appears that my TPIC8101 is not entering the Advanced SPI Mode.  I mentioned note 6 in table 2 indicating I should get an inversed response if the chip was in Advanced mode  - which I did not.

Then John suggested to another to send via spi 0x2A twice.  If the chip was in advanced mode then the second response should be 0x01, but I simply get back 0x2A once again indicating I am failing to put the chip in Advance mode.

Apparently my Microcontroller and the TPIC8101 are communicating in normal mode in that I get an echo back of the command sent, so one would think that sending 0x71 would result in the TPIC8101 going into Advanced mode, but it does not.

As I mentioned I do get the analog output when I cycle INTHOLD line, but it is about a 7 step ramp stair case from near zero up to Vdd (4.5V).  But, its clearly not being triggered by my input signal as I can remove the signal and still get the series of constant steps.

I presume there is no source code available that might have a segment of code which is know to work?

Best Regards

Ed

Hi Clancy,

I found this resolved problem on the forum:

In reply to Clancy Soehren:

Clancy

 

please disregard - the 3.3V microcontroller was strobing the SPI bus - the power supply for the TPIC was 5.26V. The signaling level was outside the 70% of VDD

for now we lowered the VDD of the TPIC to ~4.5V and it now works

is 4.5V acceptable? - the spec is not clear on a min operating voltage

It turns out my PIC32 microcontroller operates with a VDD of 3.3 volts, but I have 5 volts VDD feeding the TPIC8101.  However, it this were my problem, then I would suspect that I would get no SPI comm between my controller and the TPIC8101 - but, I do in normal mode.  I just can  not get it to go into advanced mode.  Nor does the analogy integrator appear to be responding to any signal input.

Ed

 

Excellent! I'm glad you were able to debug the issue. Please let me know if you have any questions I can help with.

Hi Clancy, yes me again!

Well, made good progress with the SPI digitized Integrator output and am getting a response.  However, it appears that the TPIC8101 is not producing the D9D8000000 response in reply to the PRESCALER command - the response to the "Channel select" command in advanced SPI mode appears normal.  

As I understand it, for a digital output from the integrator over the SPI, that the response to the PRESCALER command should be of the form D7-D0 (8bits). The response to the "Channel Select" command should be of the form D9D8000000.   Then this D9D8 value would be shifted left 2 bits and then combined with the response to the "PRESCALER" command to form the digitized voltage output value.

However, as you can see below.  The D9D8 only ever shows 255 or 0.  When it should show 64, 128 and 192.  When shifted left 2 bits they then become 256, 512 and 768 when "OR" with the D7D0 values should give the end value.  But, I never get 64, 128 or 192, its just alternates between 255 and zero.  Its clear the chip knows when it is suppose to change the D9D8 values, but it appears to simply alternate it between 255 and 0.  It changes when it should, but it does not change to the values I expected 64, 128 and 192. 

I printed the output for the responses to the PRESCALER and Channel Select commands separately so I could figure out why I was not getting the answers I expected.  The Analog voltage responses were also measure and were very close to those I calculated from the formulas in the data sheet, so it appears that part is working fine.  Its just the DA SPI response for D9D8 that does not seem to correspond to the data sheet.

The numbers code in blue are the actual values I received over the SPI.  The red numbers are the ones I expected for D9D8 but only got alternating 255 or zero.

I input varying sin wave amplitudes at a bandpass freq of 4.4Khz.  The analog values do decrease if I tune the function generator off of that center freq, in fact, everthing seems to be working correct, except the digitized ouput of the integrator.

Ts = 180usec Gain = 1 Bandpass = 4.4Khz

If you have any suggestions or if you have a TPIC8101 Test set up it would be nice to have this refuted or confirmed (for my sanity {:>)

Thanks

Ed

Hummm, seems like the jpg of the data did not come through on the previous post. I will try again

Sorry 'bout that

Ed

Ed,

In response to your e-mail:

First, the advanced mode appears to be specified correctly in the datasheet . Remember that the response is delayed by one byte, so if you are alternating between commands 1 and 2, the response will be delayed.  Maybe this example will help explain it:

Send command 1 -> receive Null

Send command 2 -> receive response 1 (D7-D0)

Send command 1 -> recieve response 2 (D9-D8)

Send command 2 -> receive response 1 (D7-D0)

Send command 2 -> receive response 2 (D9-D8)

For your comments about some data appearing too low, is this a repeatable measurement?  I did a sweep along the same area of interest as you, and did not see any problems.  Is it possible that you are sampling at the wrong time occassionally? 

Thanks,

-Clancy

Thanks, Clancy

Yes, you are correct, I am alternating between Channel 1 and 2 inputs on the chip.  Thanks for clarifying the SPI protocol, I thought I had understood it but clearly did not.  I should be able to follow your simple example.

I am going to assume that if you did not see the anomalies in SPI digitized integrator output - then it was most likely some error on my part.  If it shows up again, I will dig into it.

Thanks again for you help

Ed