Mosley PCL 6020 Mods

[chriscollins]

Back in the day, I know there used to be a company run by Greg Ogonowski who did mods to PCL series STL's to greatly reduce overshoot. Does anyone know what the modifications were? How effective were they?

 

[The Wattcher]

I think Greg took the pre EQ approach to reduce overshoot.

http://www.indexcom.com/solution/peak.html

I never used one of his modified boxes and cannot comment
on how well it worked.


I know another engineer who redesigned the AFC loop filter
to solve the overshoot problems in a PCL-606/C. I modified a couple
with his loop filter and they worked well.

The 606 is actually a better sounding box than the 6020 and much easier
to eliminate the overshoot problem in.


From Transmission Audio Processing by Robert Orban:

To ensure that the STL does not distort the shape of the audio waveform (preventing intro-
duction of overshoot into peak-limited waveforms applied to the STL input), the frequency
response must be flat (±0.1dB) throughout the operating frequency range. The group delay
must be essentially constant throughout this range (deviation from linear phase <±10°). Phase
correction can be applied to meet the requirement at high frequencies.
At low frequencies, by far the best way to achieve the specification is to extend the ¬3dB
frequency of the STL to 0.15Hz or lower and to eliminate any peaking in the infrasonic fre-
quency response prior to the rolloff. Poor AFC-loop design in STL transmitters is all too
common, and this is the most likely cause of low-frequency response problems. Such problems
can be corrected by applying equalization prior to the STL transmitter that is complementary to
existing low-frequency rolloff, such that the overall system frequency response rolls off
smoothly at 0.15Hz or below. This solution is far better than clipping the tilt-induced
overshoots after the STL receiver because the clipping will introduce non-linear distortion,
while the equalizer is distortion-free.

 

[BobOnTheJob]

How much overshoot is inherent in the 606 and 6000's?

 

[chriscollins]

I think Greg took the pre EQ approach to reduce overshoot.

http://www.indexcom.com/solution/peak.html

I never used one of his modified boxes and cannot comment
on how well it worked.


I know another engineer who redesigned the AFC loop filter
to solve the overshoot problems in a PCL-606/C. I modified a couple
with his loop filter and they worked well.

The 606 is actually a better sounding box than the 6020 and much easier
to eliminate the overshoot problem in.


From Transmission Audio Processing by Robert Orban:

To ensure that the STL does not distort the shape of the audio waveform (preventing intro-
duction of overshoot into peak-limited waveforms applied to the STL input), the frequency
response must be flat (±0.1dB) throughout the operating frequency range. The group delay
must be essentially constant throughout this range (deviation from linear phase <±10°). Phase
correction can be applied to meet the requirement at high frequencies.
At low frequencies, by far the best way to achieve the specification is to extend the ¬3dB
frequency of the STL to 0.15Hz or lower and to eliminate any peaking in the infrasonic fre-
quency response prior to the rolloff. Poor AFC-loop design in STL transmitters is all too
common, and this is the most likely cause of low-frequency response problems. Such problems
can be corrected by applying equalization prior to the STL transmitter that is complementary to
existing low-frequency rolloff, such that the overall system frequency response rolls off
smoothly at 0.15Hz or below. This solution is far better than clipping the tilt-induced
overshoots after the STL receiver because the clipping will introduce non-linear distortion,
while the equalizer is distortion-free.

We have some 606's as well... Do you still have that design? Can you share it?

 

[mm11]

From Transmission Audio Processing by Robert Orban:

To ensure that the STL does not distort the shape of the audio waveform (preventing intro-
duction of overshoot into peak-limited waveforms applied to the STL input), the frequency
response must be flat (±0.1dB) throughout the operating frequency range. The group delay
must be essentially constant throughout this range (deviation from linear phase <±10°). Phase
correction can be applied to meet the requirement at high frequencies.
At low frequencies, by far the best way to achieve the specification is to extend the ¬3dB
frequency of the STL to 0.15Hz or lower and to eliminate any peaking in the infrasonic fre-
quency response prior to the rolloff. Poor AFC-loop design in STL transmitters is all too
common, and this is the most likely cause of low-frequency response problems. Such problems
can be corrected by applying equalization prior to the STL transmitter that is complementary to
existing low-frequency rolloff, such that the overall system frequency response rolls off
smoothly at 0.15Hz or below. This solution is far better than clipping the tilt-induced
overshoots after the STL receiver because the clipping will introduce non-linear distortion,
while the equalizer is distortion-free.

I've always wondered about the AFC modifications because every exciter and STL transmitter that I have worked on uses an electrolytic blocking / coupling capacitor between the modulated oscillator and the composite baseband.
That means that the composite is going to roll of somewhere in the tens of Hz.
0.1 Hz is not going to be possible through an electrolytic capacitor that one could fit into an STL transmitter. The audio domain tests sets that I’ve used, Audio Precision and now Boonton, don’t even pass 10 Hz. So it must be that Bob and Greg are looking at the AFC loop independent of the composite baseband. It’s hard for me to grasp how that will make a difference if the sub sonic audio never even makes it to the modulated oscillator. I also wonder what they are using to measure down to 0.1Hz.

 

[WNTIRadio]

It's not the subsonic audio that's the problem, it's the peaking before the roll off. Plus, almost every filter will overshoot. That's why the filtering is so important in how it's implemented in audio processors.

 

[mm11]

It's not the subsonic audio that's the problem, it's the peaking before the roll off. Plus, almost every filter will overshoot. That's why the filtering is so important in how it's implemented in audio processors.

It appears to me that the AFC is constantly correcting that is what causes the overshoots. So I think you mean the peaking and filtering in the AFC loop only. Please clarify this for me since it's not intuitive to me.The only filtering that I see in signal path is the dominate pole of the electrolytic capacitor as a high pass filter. That's a big difference from a ringing low pass filter in a processor.

 

[Dr. Bob]

No disrespect WNTI but it is the sub-sonic audio that causes the problems, specifically the audio processor making square waves with sub-sonic components from intermodulation products caused by clipping. If the frequency and phase response of the STL is not linear where the intermod products are located, overshoot will occur. The short term AFC response also must be modified to be very slow to prevent AFC tracking the sub-sonic audio.

Having had an exciter modified by Gregg, the techniques do work but lengthen the AFC lock to over a minute. He didn't offer modifications to the 6010/6020/6030/6060 as Moseley changed the design several times and it was mostly surface mount. The 606 didn't have those issues.

Bob

 

[WNTIRadio]

Thank you for clarifying what I was trying to say! Sometimes the words just aren't there....

Yes, it is the infrasonic stuff that's caused by the audio processor, coupled with the non-linear response down that low. The "peaking" before the roll off that tilts the square waves and causes your modulation to be higher.

Same thing with AM antenna systems that have non linear response on either sideband. That's why processor manufacturers have tilt correction adjustments, so the square waves are flat.

The manual for the AM Optimods has a good explanation of this. Same applies to the STL.

 

[Dr. Bob]

Exactly WNTI, and analog FM exciters's too. One method to check the low frequency audio response is running a 10 Hz square wave through the system. If it is post processing, it needs to be flat in frequency and phase response. The non-flat frequency frequency and/or phase response will show up as tilt, overshoot, or ringing of the square waves which will also appear on the mod monitor.

Some engineers don't know about the magic of using square waves as a test signal in not only STLs or AM transmitters but in audio amplifiers like those in DAs and mixing consoles.

Bob

 

[mm11]

So how the heck can they claim that they measure the frequency response of an STL transmitter / reciever pair as flat down to below 1 hz when both of these pieces of equipment are capacitor coupled? It makes no sense.

 

[FFoti1]

So how the heck can they claim that they measure the frequency response of an STL transmitter / reciever pair as flat down to below 1 hz when both of these pieces of equipment are capacitor coupled? It makes no sense.

It depends on the size of the capacitor, and resistor values in audio circuit.

-Frank Foti

 

[Dr. Bob]

So how the heck can they claim that they measure the frequency response of an STL transmitter / reciever pair as flat down to below 1 hz when both of these pieces of equipment are capacitor coupled? It makes no sense.

It depends on the size of the capacitor, and resistor values in audio circuit.

-Frank Foti

Any basic analog circuits course text can give you theory and equations. FYI, a series 10 uF cap with a shunt resistance of 100k ohms have a 3 dB point of 0.16 hz.

Bob