Rod, at 10:16 on Mar 15 1996, you wrote to Bill Grimsley...

BG> Well, the new EPROM is certainly more reliable than the old one.  No 
BG> problems at all with the Hayes Optima now.  Have a squiz at these stats.

BG> Running a Hayes Optima 28k8 V.34 (Line 1)

BG> Running a Maestro V288FMEI 28k8 V.34 (Line 2)

RS> Interesting. 

Yeah, you could say that.  :)

RS> Interesting that the line probe stats are even worse than when you call 
RS> Pauls M34F by a tad and that all the non USR modems you have called are WAY 
RS> down on the claimed line probe db levels. Something very suspicious indeed 
RS> going on there somewhere with the USR data.

Oh well, I've put the ball back in Joe's court now.  It'll actually be
rather interesting to see how he handles what he'll likely consider to be
criticism of his beloved Couriers.  :)

RS> Unfortunately we dont appear to have any line probe stats
RS> on the calls you made to CHH before your rom change. It
RS> would be interesting to see if there was any change.

We do, as it happens.  Fortunately, I kept the logs for those two calls, so
let's have a look at them again, shall we ?

------------------------
 Hayes Optima 28k8 V.34
------------------------
Freq    Level    Level

Hertz    V.34     V.34+

 150     40       40
 300     31       30
 450     29       29
 600     29       29
 750     29       29
 900     29       29
1050     30       30
1200     31       30
1350     31       31
1500     32       32
1650     33       33
1800     34       34
1950     35       34
2100     35       35
2250     36       36
2400     37       37
2550     37       37
2700     38       38
2850     39       39
3000     40       40
3150     41       41
3300     42       42
3450     45       45
3600     52       52
3750     62       62

OK
---------------------

Although these calls were several days apart, the differences in line
levels are so insignificant as to be virtually negligible.  What the above
logs DON'T show are the respective link rates, 14400/14400/ARQ/V34/LAPM
with the original V.34 EPROM, and 28800/28800/ARQ/V34/LAPM with the V.34+
EPROM, but for the SAME FUCKING LINE LEVELS !  And if it wasn't written
above in black and white, I'm not at all convinced that I'd believe it,
either.

Unfortunately, I didn't keep the link stats for the first call (with the
V.34 EPROM), so I'll call again right now, and import the full stats below.


----------------------------------
atdt32644300
CONNECT 28800/ARQ/V34/LAPM/VB42BIS
OK
aty11
Freq     Level
 150     40
 300     30
 450     29
 600     28
 750     29
 900     29
1050     29
1200     30
1350     31
1500     32
1650     33
1800     33
1950     34
2100     35
2250     36
2400     37
2550     37
2700     38
2850     39
3000     40
3150     41
3300     42
3450     45
3600     52
3750     62

OK
ati11
USRobotics Sportster 33600 V.34+ Fax Link Diagnostics...

Modulation               V.34
Carrier Freq (Hz)        1829/1829
Symbol Rate              3200/3200
Trellis Code             64S-4D/16S-4D
Nonlinear Encoding       ON/ON
Precoding                ON/ON
Shaping                  ON/ON
Preemphasis (-dB)        6/10
Recv/Xmit Level (-dB)    35/12
Roundtrip Delay (msec)   4
Data Compression         V42BIS 2048/32
Equalization             Long
Fallback                 Enabled
Protocol                 LAPM 128/15
Speed                    26400/28800
Last Call                00:01:11
Disconnect Reason        DTR dropped
----------------------------------------

Once again, line levels appear to be pretty consistent (IOW, fairly
piss-poor), yet there was still no trouble maintaining what I'd call a damn
good connect, given the circumstances and the prevailing conditions at my
end.

Regards, Bill

--- Msgedsq/2 3.20