DRS4 Forum
  DRS4 Discussion Forum, Page 37 of 39  Not logged in ELOG logo
ID Date Author Subject
  56   Thu Mar 18 22:10:41 2010 Stefan RittSerial Interface Frequency of the DRS Chip

Hao Huan wrote:

 

 Thanks! The suggested algorithm looks promising. However, if the spikes take place only for those specific cells, is it possible to absorb them into the offset calibration?

No, since they are not constant. The bus segments charge up between readouts with a time constant of about 0.5s. So if you do the readout with 1Hz event rate and with 100Hz event rate, the peaks will differ by a factor up to 10, so a constant offset correction cannot take care of that.

  55   Thu Mar 18 21:38:10 2010 Hao HuanSerial Interface Frequency of the DRS Chip

Stefan Ritt wrote:

Hao Huan wrote:

in the DRS4 datasheet I read that the optimal frequency for SRCLK is 33MHz. However in the evaluation board firmware SRCLK is toggled at rising edges of the internal 33MHz clock, i.e. the frequency of SRCLK itself is 16.5MHz instead. Is that frequency better than 33MHz?

The reason for the 16.5 MHz is the following:

After each block of 32 bins, the DRS4 chip switches an internal segment, which causes some small spike at the analog output of the chip. This spike is a bit wider than 30ns, so if everything is digitized with 33 MHz, then you see small spiked each 32 cells. The appropriate solution would be to modify the firmware to digitize all cells with 30ns (33 MHz) and all cells with the spike with ~50 ns (20 MHz). If you do the ROI readout mode, you don't know for the first 10 cells if one of them belong to this class, since the cell address takes 10 cycles to be read out. So you would first have to read 10 cells, and then if you realize that one of them is one of the problematic ones (cell number modulo 32 is zero), you have to re-read the first 10 cells, and digitize the problematic cell with a longer settling time. Now this is a bit complicated to implement in the firmware, so I was just too lazy to do it and decided to digitize everything with 16.5 MHz. But if you are worried about the dead time, you should consider implementing the mentioned algorithm. 

 Thanks! The suggested algorithm looks promising. However, if the spikes take place only for those specific cells, is it possible to absorb them into the offset calibration?

  54   Fri Mar 12 08:04:44 2010 Stefan RittInput Bandwidth of the DRS Chip

Hao Huan wrote:

I read in the DRS datasheet that the input bandwidth if 950MHz. However, it also says the output bandwidth in the transparent mode is 50MHz. Since in the transparent mode the input is routed to the output, does it mean the input bandwidth also gets reduced in the transparent mode? I don't know how the transparent mode works inside the chip of course, but this value would be important since if the hardware discriminators are connected to the output of DRS, we have to always work in the transparent mode.

In transparent mode, the input signal also gets routed to the output, where it goes through an output buffer, which limits the bandwidth to about 50 MHz, but only for the output. The effective bandwidth to the sampling cells is not changed. Please note however that the 950 MHz are for the "chip only". We measured this by keeping the input amplitude from a function generator constant at the input pin of the chip (measured with a high speed oscilloscope). Since each signal source has a non-zero impedance, the signal tends to "shrink" at high bandwidth, and we had to adjust the level of the function generator to keep the amplitude constant at high frequencies. If you do a realistic input stage with the THS4508 for example, the achievable bandwidth will be around 800 MHz.

  53   Thu Mar 11 21:37:32 2010 Hao HuanInput Bandwidth of the DRS Chip

Hi Stefan,

    I read in the DRS datasheet that the input bandwidth if 950MHz. However, it also says the output bandwidth in the transparent mode is 50MHz. Since in the transparent mode the input is routed to the output, does it mean the input bandwidth also gets reduced in the transparent mode? I don't know how the transparent mode works inside the chip of course, but this value would be important since if the hardware discriminators are connected to the output of DRS, we have to always work in the transparent mode.

    Thanks!

 

  52   Thu Mar 11 11:45:52 2010 Stefan RittReadout of DRS Data

Hao Huan wrote:

Hi Stefan,

    thanks to your help I can now successfully keep the Domino wave running at a stable frequency and maintain the channel cascading information in the Write Shift Register. (Since you told me WSR always reads and writes at the same time, I think I need to rewrite the information back every time after reading out from WSR to decide from which channel my data come, don't I?)

Yes you do. But if you have WSRLOOP=1 in the config register, this is done automatically. So the SR output is visible at the pin and will be fed back into the input.

Hao Huan wrote:

    However I'm still having difficulty in reading out from the DRS cells. I use the ROI readout mode and assume as long as I give a pulse on RSRLOAD the data will come out one by one. 

That's not correct. Have a look at Figure 14 of the datasheet. Do you see a single RSRLOAD pulse or many? There is only one RSRLOAD pulse to initialize the readout shift register, then the cells are clocked by SRCLK pulses. 

Hao Huan wrote:

Also I read in the datasheet that WSROUT will give RSR output when DWRITE is low. Sometimes I see some random bits from this output and sometimes I see all zero's. What is the reasonable output I should see from WSROUT, say, when I'm running in the transparent mode with DWRITE low? 

A single RSRLOAD pulse loads the RSR with a "1" at the domino stop position and "0" in all other places. A pulse on SRCLK shifts this "1" down the RSR. When it arrives at cell #1023, it will be visible for one clock cycle at WSROUT. The "double" functionality of WSROUT has the following background: Assume you use channel cascading 2x2048. Now the domino wave stopps in cell 1020 of the first channel for example. You have to read cells 1020,1021,1022,1024 of the first channel, then you continue with 0,1,2 on the second channel. But how do you know that you have to switch channels after the first four clock cycles? The SROUT output encodes the stop position (in this case 1020), but it needs 10 clock cycles before the information is available, so you don't have it after four cycles. That's where WSROUT comes into play: Since it outputs RSR bit by bit, it will show three "0", then a "1", when you are at cell 1023. Then you know that you have to switch channels immediately. That's why I output RSR via WSROUT if DWRITE is low.

 

  51   Wed Mar 10 10:07:28 2010 Stefan RittSerial Interface Frequency of the DRS Chip

Hao Huan wrote:

in the DRS4 datasheet I read that the optimal frequency for SRCLK is 33MHz. However in the evaluation board firmware SRCLK is toggled at rising edges of the internal 33MHz clock, i.e. the frequency of SRCLK itself is 16.5MHz instead. Is that frequency better than 33MHz?

The reason for the 16.5 MHz is the following:

After each block of 32 bins, the DRS4 chip switches an internal segment, which causes some small spike at the analog output of the chip. This spike is a bit wider than 30ns, so if everything is digitized with 33 MHz, then you see small spiked each 32 cells. The appropriate solution would be to modify the firmware to digitize all cells with 30ns (33 MHz) and all cells with the spike with ~50 ns (20 MHz). If you do the ROI readout mode, you don't know for the first 10 cells if one of them belong to this class, since the cell address takes 10 cycles to be read out. So you would first have to read 10 cells, and then if you realize that one of them is one of the problematic ones (cell number modulo 32 is zero), you have to re-read the first 10 cells, and digitize the problematic cell with a longer settling time. Now this is a bit complicated to implement in the firmware, so I was just too lazy to do it and decided to digitize everything with 16.5 MHz. But if you are worried about the dead time, you should consider implementing the mentioned algorithm. 

  50   Tue Mar 9 23:28:45 2010 Hao HuanSerial Interface Frequency of the DRS Chip

Hi Stefan,

    in the DRS4 datasheet I read that the optimal frequency for SRCLK is 33MHz. However in the evaluation board firmware SRCLK is toggled at rising edges of the internal 33MHz clock, i.e. the frequency of SRCLK itself is 16.5MHz instead. Is that frequency better than 33MHz?

    Thanks!

 

  49   Fri Mar 5 23:29:04 2010 Hao HuanReadout of DRS Data

Hao Huan wrote:

Hi Stefan,

    thanks to your help I can now successfully keep the Domino wave running at a stable frequency and maintain the channel cascading information in the Write Shift Register. (Since you told me WSR always reads and writes at the same time, I think I need to rewrite the information back every time after reading out from WSR to decide from which channel my data come, don't I?)

    However I'm still having difficulty in reading out from the DRS cells. I use the ROI readout mode and assume as long as I give a pulse on RSRLOAD the data will come out one by one. However, what I get is just a constant with some noise, which seems I'm always reading from the same cell. Actually I'm not very clear about how it works. What's the mechanism for RSRLOAD  and do I have to initialize the Read Shift Register to use the ROI mode? Also I read in the datasheet that WSROUT will give RSR output when DWRITE is low. Sometimes I see some random bits from this output and sometimes I see all zero's. What is the reasonable output I should see from WSROUT, say, when I'm running in the transparent mode with DWRITE low?

    Thank you very much!

 

Hi Stefan,

    I tried again and confirmed the problem... In the full readout mode I could successfully read out all the data, but in the ROI mode if I naively apply a pulse at RSRLOAD the results are not right. Is there anything I should be careful about in the ROI readout mode?

    Thanks!

 

  48   Thu Mar 4 19:14:10 2010 Hao HuanReadout of DRS Data

Hi Stefan,

    thanks to your help I can now successfully keep the Domino wave running at a stable frequency and maintain the channel cascading information in the Write Shift Register. (Since you told me WSR always reads and writes at the same time, I think I need to rewrite the information back every time after reading out from WSR to decide from which channel my data come, don't I?)

    However I'm still having difficulty in reading out from the DRS cells. I use the ROI readout mode and assume as long as I give a pulse on RSRLOAD the data will come out one by one. However, what I get is just a constant with some noise, which seems I'm always reading from the same cell. Actually I'm not very clear about how it works. What's the mechanism for RSRLOAD  and do I have to initialize the Read Shift Register to use the ROI mode? Also I read in the datasheet that WSROUT will give RSR output when DWRITE is low. Sometimes I see some random bits from this output and sometimes I see all zero's. What is the reasonable output I should see from WSROUT, say, when I'm running in the transparent mode with DWRITE low?

    Thank you very much!

 

  47   Wed Mar 3 17:49:30 2010 Stefan RittInitialization of the Domino Circuit

Hao Huan wrote:

Hi Stefan,

    I read in the datasheet that every time after power up the Domino wave in DRS4 needs to be started and stopped once to initialize the Domino circuit. However in your firmware it seems the chip immediately goes into the idle state after reset. Is that Domino circuit initialization really necessary?

    Also an aside question: in your firmware the readout process has the SRCLK sent to DRS4 only about 200ns later after RSRLOAD gets asserted instead of immediately following RSRLOAD. Is there any reason for that?

    Thanks a lot!

 

The start/stop requirement is obsolete and has been replaced by  elog:10. I need to update this in the datasheet. The delay between the RSRLOAD and the SRCLK has the following reason: On the RSRLOAD the first sampling cell is output to the chip and to the ADC. This can sometimes be a rather high swing, which needs some time to settle, and some warm-up for the output driver. But actually I never really measured it, so it's there just as a safety margin. But I would encourage you to try to reduce this time and see it the first few bins of the readout change in offset.

  46   Wed Mar 3 17:36:31 2010 Hao HuanInitialization of the Domino Circuit

Hi Stefan,

    I read in the datasheet that every time after power up the Domino wave in DRS4 needs to be started and stopped once to initialize the Domino circuit. However in your firmware it seems the chip immediately goes into the idle state after reset. Is that Domino circuit initialization really necessary?

    Also an aside question: in your firmware the readout process has the SRCLK sent to DRS4 only about 200ns later after RSRLOAD gets asserted instead of immediately following RSRLOAD. Is there any reason for that?

    Thanks a lot!

 

  45   Wed Mar 3 14:37:40 2010 Stefan RittPLLLCK signal of DRS4

Hao Huan wrote:

Stefan Ritt wrote:

Hao Huan wrote:

By the way I have another question: when the default operation mode of the DRS4 chip is used, i.e. WSRIN is fed internally to WSROUT, at the external pins is it necessary to leave the WSRIN open? Or any input through the pin will not affect the Domino wave running? Also I observe that WSROUT will be always low when the chip is running in this mode; is it the supposed behavior?

If the WSRin is fed internally to WSROUT, then the level of the WSRIN pin does not matter, it's just disconnected. You can leave the pin open without problem. WSROUT is however active, so you can observe the internal state of the write shift register. In the default configuration (8x1024 sampling cells), all 8 channels are active all the time, so the WSR is loaded with ones. The inverter at the output then makes all zeros from this. If you configure the chip as 4x2048 cells, then you will observe switching bits at WSROUT. 

 Sorry; WSROUT also has an inverter? Actually I have one more stupid question about the shift registers: when we assert the address bits to operate on one shift register, e.g. WSR, we use SRIN to give input and SROUT to read output; but how does the shift register know whether we're reading or writing? Or it will just receive input from SRIN and give output at SROUT at the same time?

Actually I double checked the schematics, WSROUT has NO inverter at the output. So the output should be always one in a 8x1024 channel configuration.

Concerning the read/write you are right. On each clock cycle, SRIN will be shifted into the first bit, and the last bit will be visible at SROUT.

  44   Mon Feb 22 17:23:59 2010 Hao HuanPLLLCK signal of DRS4

Stefan Ritt wrote:

Hao Huan wrote:

By the way I have another question: when the default operation mode of the DRS4 chip is used, i.e. WSRIN is fed internally to WSROUT, at the external pins is it necessary to leave the WSRIN open? Or any input through the pin will not affect the Domino wave running? Also I observe that WSROUT will be always low when the chip is running in this mode; is it the supposed behavior?

If the WSRin is fed internally to WSROUT, then the level of the WSRIN pin does not matter, it's just disconnected. You can leave the pin open without problem. WSROUT is however active, so you can observe the internal state of the write shift register. In the default configuration (8x1024 sampling cells), all 8 channels are active all the time, so the WSR is loaded with ones. The inverter at the output then makes all zeros from this. If you configure the chip as 4x2048 cells, then you will observe switching bits at WSROUT. 

 Sorry; WSROUT also has an inverter? Actually I have one more stupid question about the shift registers: when we assert the address bits to operate on one shift register, e.g. WSR, we use SRIN to give input and SROUT to read output; but how does the shift register know whether we're reading or writing? Or it will just receive input from SRIN and give output at SROUT at the same time?

Thank you so much!

  43   Sun Feb 21 20:33:57 2010 Stefan RittPLLLCK signal of DRS4

Hao Huan wrote:

By the way I have another question: when the default operation mode of the DRS4 chip is used, i.e. WSRIN is fed internally to WSROUT, at the external pins is it necessary to leave the WSRIN open? Or any input through the pin will not affect the Domino wave running? Also I observe that WSROUT will be always low when the chip is running in this mode; is it the supposed behavior?

If the WSRin is fed internally to WSROUT, then the level of the WSRIN pin does not matter, it's just disconnected. You can leave the pin open without problem. WSROUT is however active, so you can observe the internal state of the write shift register. In the default configuration (8x1024 sampling cells), all 8 channels are active all the time, so the WSR is loaded with ones. The inverter at the output then makes all zeros from this. If you configure the chip as 4x2048 cells, then you will observe switching bits at WSROUT. 

  42   Sun Feb 21 20:27:46 2010 Hao HuanPLLLCK signal of DRS4

Stefan Ritt wrote:

Hao Huan wrote:


Thanks! I see. The capacitor is important. However I'm a little confused... If PLLLCK=DTAP XOR REFCLK, shouldn't it integrate to low instead of high when the two clocks are in phase? I must have some misunderstanding here. So if we ignore any realistic complexity and assume DTAP is perfectly locked with REFCLK, will PLLLCK be always low or high? I'm sorry I do not know how the DRS internal PLL and its input/output work...

Actually the XOR is followed by an inverter, so it will integrate to high if the two clocks are in phase. 

 Got it. Thank you! By the way I have another question: when the default operation mode of the DRS4 chip is used, i.e. WSRIN is fed internally to WSROUT, at the external pins is it necessary to leave the WSRIN open? Or any input through the pin will not affect the Domino wave running? Also I observe that WSROUT will be always low when the chip is running in this mode; is it the supposed behavior?

  41   Sun Feb 21 13:47:03 2010 Stefan RittPLLLCK signal of DRS4

Hao Huan wrote:


Thanks! I see. The capacitor is important. However I'm a little confused... If PLLLCK=DTAP XOR REFCLK, shouldn't it integrate to low instead of high when the two clocks are in phase? I must have some misunderstanding here. So if we ignore any realistic complexity and assume DTAP is perfectly locked with REFCLK, will PLLLCK be always low or high? I'm sorry I do not know how the DRS internal PLL and its input/output work...

Actually the XOR is followed by an inverter, so it will integrate to high if the two clocks are in phase. 

  40   Sun Feb 21 13:41:35 2010 Stefan RittReal Time Conference 2010

Hello,

may I draw your attention to the upcoming Real Time Conference 2010, taking place in Lisbon, Portugal, May 23rd to May 28th, 2010.

http://rt2010.ipfn.ist.utl.pt/

Several groups which are developing DRS4 electronics will come to this conference and present their work, so it will be a good opportunity to exchange ideas and experiences. There will also be a short course on digital pulse shape processing, which is highly relevant for our field.

Looking forward to see you in Lisbon,

    Stefan 

  39   Sun Feb 21 00:46:01 2010 Hao HuanPLLLCK signal of DRS4

Stefan Ritt wrote:

Hao Huan wrote:

Hi Stefan,

    in the latest DRS4 datasheet I only saw your data of the DRS4 PLL locking time for 6GSPS sampling speed, with other rows "TBD". Have you tried those lower frequencies? According to the datasheet I think the PLLLCK should be stabily low when the PLL is locked; am I right? However when I try my design with the DRS4 chip and feed the reference clock signal at 0.5MHz or 2MHz, the PLLLCK I get can never stabilize. There could be some problem in the PCB circuit connection, but I want to confirm with you since I'm confused with those "TBD" blanks.

    Thanks a lot!

 

The locking time is typically 20-30 cycles of the external reference clock, I will update the numbers in the datasheet soon. I attached a screenshot of the chip when starting up at 1 GHz (0.5 MHz REFCLK), so you can see the behaviour. The upper curver is the DTAP signal, the lower curve the PLLLCK signal. As you can see, the PLLLCK signal is not purely digital. Actually it's a simple XOR between the REFCLK and the DTAP signal, so you need an external 4.7nF capacitor to "integrate" this signal. Without this capacitor, you would see small negative spikes whenever there is s small phase shift between the DTAP and the REFCLK signal. Have a look at your DTAP signal, is it in phase with the REFCLK? 

Thanks! I see. The capacitor is important. However I'm a little confused... If PLLLCK=DTAP XOR REFCLK, shouldn't it integrate to low instead of high when the two clocks are in phase? I must have some misunderstanding here. So if we ignore any realistic complexity and assume DTAP is perfectly locked with REFCLK, will PLLLCK be always low or high? I'm sorry I do not know how the DRS internal PLL and its input/output work...

  38   Sat Feb 20 09:54:48 2010 Stefan RittPLLLCK signal of DRS4

Hao Huan wrote:

Hi Stefan,

    in the latest DRS4 datasheet I only saw your data of the DRS4 PLL locking time for 6GSPS sampling speed, with other rows "TBD". Have you tried those lower frequencies? According to the datasheet I think the PLLLCK should be stabily low when the PLL is locked; am I right? However when I try my design with the DRS4 chip and feed the reference clock signal at 0.5MHz or 2MHz, the PLLLCK I get can never stabilize. There could be some problem in the PCB circuit connection, but I want to confirm with you since I'm confused with those "TBD" blanks.

    Thanks a lot!

 

The locking time is typically 20-30 cycles of the external reference clock, I will update the numbers in the datasheet soon. I attached a screenshot of the chip when starting up at 1 GHz (0.5 MHz REFCLK), so you can see the behaviour. The upper curver is the DTAP signal, the lower curve the PLLLCK signal. As you can see, the PLLLCK signal is not purely digital. Actually it's a simple XOR between the REFCLK and the DTAP signal, so you need an external 4.7nF capacitor to "integrate" this signal. Without this capacitor, you would see small negative spikes whenever there is s small phase shift between the DTAP and the REFCLK signal. Have a look at your DTAP signal, is it in phase with the REFCLK? 

  37   Sat Feb 20 01:56:05 2010 Hao HuanPLLLCK signal of DRS4

Hi Stefan,

    in the latest DRS4 datasheet I only saw your data of the DRS4 PLL locking time for 6GSPS sampling speed, with other rows "TBD". Have you tried those lower frequencies? According to the datasheet I think the PLLLCK should be stabily low when the PLL is locked; am I right? However when I try my design with the DRS4 chip and feed the reference clock signal at 0.5MHz or 2MHz, the PLLLCK I get can never stabilize. There could be some problem in the PCB circuit connection, but I want to confirm with you since I'm confused with those "TBD" blanks.

    Thanks a lot!

 

ELOG V3.1.4-80633ba