Log entry time 09:18:26 on August 2,2001

Entry number 68603

This entry is a followup to: 68573

Followups:

• 68612 08/02/01 11:50 gilman,reitz g2 helicity setup

Date: Thu, 2 Aug 2001 07:26:47 -0400 (EDT)

From: Ronald Gilman 

To: Robert Michaels 

Cc: jpchen@jlab.org, averett@jlab.org, reitz@jlab.org,

     Wolfgang Korsch , Ronald Gilman 

Subject: Re: sign reversed

 

Hi Bob...

 

we raced through the setup last night; it only took Bodo and me

about 30 minutes to set up the "g2n random" mode. Here are notes

and at the end a suggestion for improving the circuit.

 

0) Important: if the helicity frequency is changed upstairs

   the circuit timing needs to be adjusted.

 

1) The glitches should not matter. Since the clock pulse is

   10 - 13 us after the "delayed helicity" transition, and the

   glitch is ~6 us after, it does not affect the sequence of

   events started by the clock.

 

2) The clock pulse has the same time variations as the

   "delayed helicity", ~ +/- 50 us full width out of 33 ms.

 

3) In the counting house:

   a) We use clock to start two channels LeCroy G&D set to

      500 us (systematic difference < 100 ns, individual

      variations +/- ~100 ns, 100 ns FW sort of number based

      on few minutes with scope, may be, eg, daily fluctuations

      due to temperature variations, etc., etc., etc.)

   b) We used the delay output of each to start LeCroy G&D

      set to ~32.7 ms, accuracy ~ 30 us I think. These second

      LeCroy channels have the "helicity +" and "helicity -" =

      "helicity + complement" inputs as inhibits, so one only

      fires on h+ and the other only fires on h-. Note again

      there should be short 10-15 ns spikes output in the counting

      house when gated by the other phase, but these will be dispersed

      away.

      - The inhibiting here is likely where I reversed the phases of +/-

        I'll check after we speak this morning.

   c) we saw that the typical time difference between end of one gate

        being TRUE and start of next (either same or other phase)

      being TRUE was typically ~650 us, which seems a reasonable safe

      choice, about as we expected, but some times it appeared to jump

      to ~800-900 us. We did not look into this, it seems too much.

      Another task for today.

 

4) We send the signals to each spectrometer. In the spectrometer,

   there is a logical fanout of the incoming signals. We use the

   start of it to start a LeCroy G&D latch, the complement output

   of the fanout to stop the latch at the end of the signal, and

   use the run gate to inhibit the latch.

   - When the run gate goes away, you will still get these 10-15 ns

     glitches from the LeCroy, which could allow some scaler counts,

     could allow the helicity gate itself to be counted, though probably

     the timing is a little wrong since the gate is so short. We probably

     should switch to Phillips. I think it is more important in the

     spectrometer than upstairs.

 

Changes: The setup downstairs suggests we can make the time in the two

phases more symmetric if we do the following to the crate upstairs.

 

1) Remove the discriminator, add in a logic fan in / fan out (we need the

   extra channels).

2) Use G&D channel a to set up 500 us delay. Use the delayed output to

   start...

3) G&D channel b, set up 32.7 ms. Put NIM output to fanout to start...

4) G&D channels c and d, which are inhibited by h+ and h-, respectively,

   and stopped by the complement of channel b output.

5) Fan out channel c and d outputs to send downstairs.

 

This way there should only be statistical fluctuations in the widths of

the two helicity phases, not systematic differences.

 

So... although you have done these checks, I suggest we switch to the

Phillips modules downstairs, and redo the crate / LeCroy circuit

upstairs as I described.

 

rg                                                                                     

On Wed, 1 Aug 2001, Robert Michaels wrote:

 

> From a short run tonight, in "g2n random" mode (our

> running config), it looks like the sign of the spectrometer

> DAQ asymmetry is reversed from before.  I suggest it be

> the same as before.  At least one should attempt to under-

> stand from the cabling why it is reversed.  I don't have

> enough statistics to comment on the agreement of magnitude

> yet.  Will tomorrow.