- Joined
- Jun 12, 2010
- Messages
- 12,037
- Location
- OZONE ALLEY MARS (Visitor)
- Tagline
- Wasted Days and Wasted Nights
Jerry's " ONLY ONE IN THE COUNTRY" Phase Linear 700 Thread
- Thread starter laatsch55
- Start date
- Joined
- Jan 14, 2011
- Messages
- 74,943
- Location
- Gillette, Wyo.
- Tagline
- Halfbiass...Electron Herder and Backass Woof
Re: Jerry's " ONLY ONE IN THE COUNTRY" Phase Linear 700 Thre
Sure was, inactive at the moment and don't give me any shit!!!!! You old curmudgeon!!!! The 3300 is a better pre anyway, doesn't have the inheren't problems the 4000 had. the 4000 can be good , just gonna take a lot of work. You oughta see the service bulletins on that begger, shoulda recalled all of em.
Sure was, inactive at the moment and don't give me any shit!!!!! You old curmudgeon!!!! The 3300 is a better pre anyway, doesn't have the inheren't problems the 4000 had. the 4000 can be good , just gonna take a lot of work. You oughta see the service bulletins on that begger, shoulda recalled all of em.
- Joined
- Jun 12, 2010
- Messages
- 12,037
- Location
- OZONE ALLEY MARS (Visitor)
- Tagline
- Wasted Days and Wasted Nights
- Joined
- Jan 14, 2011
- Messages
- 74,943
- Location
- Gillette, Wyo.
- Tagline
- Halfbiass...Electron Herder and Backass Woof
Re: Jerry's " ONLY ONE IN THE COUNTRY" Phase Linear 700 Thre
Beins Joe solved the thump problem we have converted back to the board for the + 20 volt feed for the DCprotect relays. The additional heatsink was required as the relays draw sufficient current that it swamped the original heatsink surface Joe had built in the board.
Beins Joe solved the thump problem we have converted back to the board for the + 20 volt feed for the DCprotect relays. The additional heatsink was required as the relays draw sufficient current that it swamped the original heatsink surface Joe had built in the board.
Attachments
- Joined
- Jan 14, 2011
- Messages
- 74,943
- Location
- Gillette, Wyo.
- Tagline
- Halfbiass...Electron Herder and Backass Woof
- Joined
- Jan 14, 2011
- Messages
- 74,943
- Location
- Gillette, Wyo.
- Tagline
- Halfbiass...Electron Herder and Backass Woof
Re: Jerry's " ONLY ONE IN THE COUNTRY" Phase Linear 700 Thre
IF the pre-amp is turned on before the amp is turned on NO!! If the amp is on and stable(takes a few milliseconds) and then the pre-amp is YES!! An amp will amplify whatever is present at it's input, provided that signal is within its frequency response range, in this case 3HZ to 80KHZ.
IF the pre-amp is turned on before the amp is turned on NO!! If the amp is on and stable(takes a few milliseconds) and then the pre-amp is YES!! An amp will amplify whatever is present at it's input, provided that signal is within its frequency response range, in this case 3HZ to 80KHZ.
Re: Jerry's " ONLY ONE IN THE COUNTRY" Phase Linear 700 Thre
Description of the latch up possibility.
When power is applied to the PL amp, the +/- bulk caps charge rapidly to provide the main power for the amp. During this start up phase, the critical ‘front end’ long-tailed pair formed by the differential amp (Q1 and Q2) is not in control yet because it is not provided sufficient voltage to properly bias itself and operate as an op amp. In addition, the positive and negative main rails do not rise at exactly the same rate leading to bias imbalance in this front end during this critical start up period.
Before the front end can control the amp output using global negative feedback provided from the output through R15, R13 and C6, the tendency is for the amp to come up with a positive bias on the output. This is because the positive rail feeds directly through R17 then through R18 then through the base emitter junction of Q10, turning it on quite hard on power up which in turn drives Q11, Q13, Q15, Q17 to turn on quite hard. This DC voltage is fed back through R15, R13 to begin charging C6 in the positive direction. During this startup phase, Q5 is not active yet (off) because the front end is not biased properly and cannot control the second stage current amplifier formed by Q3 and Q4 to provide base current to control Q5.
If the amp starts properly and the front end comes into regulation fast enough, the global negative feedback compensates for this and through the action of the differential front end turns on Q4 and in turn Q5 to bring the positive turn on bump back to ground again and all is well.
In the latch up condition, this start up process begins but then runs away as follows. Q10, Q11, Q13, Q15, Q17 all turn on hard due to the bias from R17 and R18. This positive bump produces a current in R15 and R13 starting to slowly charge C6. The rapid voltage rise at the nodal junction of R15 and R13 starts a current flowing through R10 and C4 into the base of Q2. This base current due to both +/- rails starting up at similar rates will flow through both the base-emitter junction (into R6) as well as the base-collector junction (into R7). The split of current flow is somewhat indeterminant but the base-collector junction is typically a few millivolts lower than the base-emitter junction in a TIS97 transistor. As a result, the tendency will be for this feedback current to flow through the path of least resistance, the base-collector junction and into R7. This is not the intent by design. This action produces a cut off of Q4, the MPSA93 transistor which continues to hold off Q5 which reinforces the continued rise of the output voltage in the positive direction. Essentially negative feedback turns into positive feedback by accident, the amp latches up and the output rail goes to the positive rail. In the case of the PL20 board with the +/-20V regulators for the front end, once the amp is latched up, the collector of Q2 rises to approximately 29V turning off Q4 very hard, giving the amp no chance of coming back into regulation. This collector node of Q2 is nominally supposed to operate at 10V. If the amp is latched up and then power cycled rapidly again, C6 will have retained much of this ~29V charge causing the amp to almost guarantee to latch up again.
How do the back to back diodes around C6 fix this?
It is mainly the diode with its anode connected to the positive side of C6 and its cathode connected to the negative side of C6 that does the trick. The other diode is added for symmetry even though the amp has no propensity to latch up in the negative direction. This first diode clamps the node at the junction of R15 and R13 to a maximum of +/- 2.36V for the PL400 or +/- 2.85V for the PL700B. This allows the front end to become properly powered and stabilized and keep the negative feedback signal within the allowable common mode voltage range of this front end stage. Any tendency to latch up will be quickly remedied once the +/- 20V supplies come into regulation. These diodes do not come into play when the amp is amplifying audible frequencies as the impedance of C6 is so low (for example, at 20Hz its impedance is only 79 ohms and will only experience 0.49V peak when the amp is producing a 80V peak signal output which is less than the Vf of the diodes involved, at 40Hz there is only 0.25V peak across C6 and so on). This is an elegant and effective fix and allows the amp to be power cycled rapidly with no concern of latch up.
What about turn on and turn off thump? These are actually artifacts of the front end not being in control during turn on and turn off. If these thumps exist, it is evidence that the amp is starting to latch up and then recover. This fix allows the negative feedback signal to more rapidly work with the front end to stabilize and regulate the output by staying well within the common mode range of the front end.
Sorry for the long post.
Joe
Description of the latch up possibility.
When power is applied to the PL amp, the +/- bulk caps charge rapidly to provide the main power for the amp. During this start up phase, the critical ‘front end’ long-tailed pair formed by the differential amp (Q1 and Q2) is not in control yet because it is not provided sufficient voltage to properly bias itself and operate as an op amp. In addition, the positive and negative main rails do not rise at exactly the same rate leading to bias imbalance in this front end during this critical start up period.
Before the front end can control the amp output using global negative feedback provided from the output through R15, R13 and C6, the tendency is for the amp to come up with a positive bias on the output. This is because the positive rail feeds directly through R17 then through R18 then through the base emitter junction of Q10, turning it on quite hard on power up which in turn drives Q11, Q13, Q15, Q17 to turn on quite hard. This DC voltage is fed back through R15, R13 to begin charging C6 in the positive direction. During this startup phase, Q5 is not active yet (off) because the front end is not biased properly and cannot control the second stage current amplifier formed by Q3 and Q4 to provide base current to control Q5.
If the amp starts properly and the front end comes into regulation fast enough, the global negative feedback compensates for this and through the action of the differential front end turns on Q4 and in turn Q5 to bring the positive turn on bump back to ground again and all is well.
In the latch up condition, this start up process begins but then runs away as follows. Q10, Q11, Q13, Q15, Q17 all turn on hard due to the bias from R17 and R18. This positive bump produces a current in R15 and R13 starting to slowly charge C6. The rapid voltage rise at the nodal junction of R15 and R13 starts a current flowing through R10 and C4 into the base of Q2. This base current due to both +/- rails starting up at similar rates will flow through both the base-emitter junction (into R6) as well as the base-collector junction (into R7). The split of current flow is somewhat indeterminant but the base-collector junction is typically a few millivolts lower than the base-emitter junction in a TIS97 transistor. As a result, the tendency will be for this feedback current to flow through the path of least resistance, the base-collector junction and into R7. This is not the intent by design. This action produces a cut off of Q4, the MPSA93 transistor which continues to hold off Q5 which reinforces the continued rise of the output voltage in the positive direction. Essentially negative feedback turns into positive feedback by accident, the amp latches up and the output rail goes to the positive rail. In the case of the PL20 board with the +/-20V regulators for the front end, once the amp is latched up, the collector of Q2 rises to approximately 29V turning off Q4 very hard, giving the amp no chance of coming back into regulation. This collector node of Q2 is nominally supposed to operate at 10V. If the amp is latched up and then power cycled rapidly again, C6 will have retained much of this ~29V charge causing the amp to almost guarantee to latch up again.
How do the back to back diodes around C6 fix this?
It is mainly the diode with its anode connected to the positive side of C6 and its cathode connected to the negative side of C6 that does the trick. The other diode is added for symmetry even though the amp has no propensity to latch up in the negative direction. This first diode clamps the node at the junction of R15 and R13 to a maximum of +/- 2.36V for the PL400 or +/- 2.85V for the PL700B. This allows the front end to become properly powered and stabilized and keep the negative feedback signal within the allowable common mode voltage range of this front end stage. Any tendency to latch up will be quickly remedied once the +/- 20V supplies come into regulation. These diodes do not come into play when the amp is amplifying audible frequencies as the impedance of C6 is so low (for example, at 20Hz its impedance is only 79 ohms and will only experience 0.49V peak when the amp is producing a 80V peak signal output which is less than the Vf of the diodes involved, at 40Hz there is only 0.25V peak across C6 and so on). This is an elegant and effective fix and allows the amp to be power cycled rapidly with no concern of latch up.
What about turn on and turn off thump? These are actually artifacts of the front end not being in control during turn on and turn off. If these thumps exist, it is evidence that the amp is starting to latch up and then recover. This fix allows the negative feedback signal to more rapidly work with the front end to stabilize and regulate the output by staying well within the common mode range of the front end.
Sorry for the long post.
Joe
Joe
- Joined
- Jan 14, 2011
- Messages
- 74,943
- Location
- Gillette, Wyo.
- Tagline
- Halfbiass...Electron Herder and Backass Woof
Re: Jerry's " ONLY ONE IN THE COUNTRY" Phase Linear 700 Thre
Joe------in no way , shape or form should you be apologizing for a long post when it was so eloquently related in that way. My first thought was " I'll be damned", not because I
understood it all, but that I understood enough to know how sim-ply elgant the fix was.
So it was dependent enough on the actual Q1-Q2 characteristics that even a small mismatch or different production runs of the same said TIS-97 could produce these same results in this inconsistent manner?? I realize that may be an oversimplification, but I'm trying to wrap my head around the intermittent factor and i believe you've answered that question.
In the original design was there enough tolerance in C6 to make this unpredictable and this particular circuit not understood to the point it is today given the help of CAD etc??
Joe------in no way , shape or form should you be apologizing for a long post when it was so eloquently related in that way. My first thought was " I'll be damned", not because I
understood it all, but that I understood enough to know how sim-ply elgant the fix was.
So it was dependent enough on the actual Q1-Q2 characteristics that even a small mismatch or different production runs of the same said TIS-97 could produce these same results in this inconsistent manner?? I realize that may be an oversimplification, but I'm trying to wrap my head around the intermittent factor and i believe you've answered that question.
In the original design was there enough tolerance in C6 to make this unpredictable and this particular circuit not understood to the point it is today given the help of CAD etc??
- Joined
- Jun 11, 2010
- Messages
- 6,850
- Location
- Powhatan, Virginia, United States
- Tagline
- WassupYa Mang?
- Joined
- Jan 14, 2011
- Messages
- 74,943
- Location
- Gillette, Wyo.
- Tagline
- Halfbiass...Electron Herder and Backass Woof
- Joined
- Jan 14, 2011
- Messages
- 74,943
- Location
- Gillette, Wyo.
- Tagline
- Halfbiass...Electron Herder and Backass Woof
Re: Jerry's " ONLY ONE IN THE COUNTRY" Phase Linear 700 Thre
Hi Lee
As with most marginally stable circuit conditions, they can fall over in the direction of stable or unstable based on a lot of factors. The B-E vs. B-C forward characteristics is one factor, this one tends toward instability due to the tendency for the B-C junction to have a slightly lower Vforward but that could vary based on production batches, temperature of the amp, etc. Another factor is how fast the voltage sources for the front end rise in relation to the voltage sources for the back end. Ideally in an amp of this construction, if the voltage source for the front end was powered up and stable (timing wise) before the back end turned on, this amp would start properly every time, no doubt about that. The fact is exactly the opposite, the front end lags the back end (timing wise) from the voltage sequencing perspective. IMO, the worst case is if the amp has been sitting for a long time with no power applied and every capacitor fully discharged or being powered up for the first time ever. The condition of C3 and C23 factor into this timing race condition, the larger they are, the better filtering they provide but at the same time they slow down the voltage startup on the front end, increasing the latch up potential. Phase Linear was onto some of this. In the original 700B service manual, on page 26 they discuss Turn Off Noise and how to affect extremely loud turn off thump by replacing D3 and C3 in the defective channel. Seems like Phase Linear was onto what factors affected it back in February of 1981 but not what it took to cure it. I have attached the relevant service manual page to this post.
Hope this helps out. If any questions, just let me know.
Joe
Hi Lee
As with most marginally stable circuit conditions, they can fall over in the direction of stable or unstable based on a lot of factors. The B-E vs. B-C forward characteristics is one factor, this one tends toward instability due to the tendency for the B-C junction to have a slightly lower Vforward but that could vary based on production batches, temperature of the amp, etc. Another factor is how fast the voltage sources for the front end rise in relation to the voltage sources for the back end. Ideally in an amp of this construction, if the voltage source for the front end was powered up and stable (timing wise) before the back end turned on, this amp would start properly every time, no doubt about that. The fact is exactly the opposite, the front end lags the back end (timing wise) from the voltage sequencing perspective. IMO, the worst case is if the amp has been sitting for a long time with no power applied and every capacitor fully discharged or being powered up for the first time ever. The condition of C3 and C23 factor into this timing race condition, the larger they are, the better filtering they provide but at the same time they slow down the voltage startup on the front end, increasing the latch up potential. Phase Linear was onto some of this. In the original 700B service manual, on page 26 they discuss Turn Off Noise and how to affect extremely loud turn off thump by replacing D3 and C3 in the defective channel. Seems like Phase Linear was onto what factors affected it back in February of 1981 but not what it took to cure it. I have attached the relevant service manual page to this post.
Hope this helps out. If any questions, just let me know.
Joe
Attachments
- Joined
- Jan 14, 2011
- Messages
- 74,943
- Location
- Gillette, Wyo.
- Tagline
- Halfbiass...Electron Herder and Backass Woof
Re: Jerry's " ONLY ONE IN THE COUNTRY" Phase Linear 700 Thre
Joe, i was also going to reference that page as proof they understood it could happen, just not what it took to clear it. Soooooo, when i powered up the 700 for the first time with your new board and virgin caps, that was the worst case scenario and the worst case happened. Drove me nuts there for a bit because it went away after trying it a couple more times, as i noted in my post when it happened.
Joe, i was also going to reference that page as proof they understood it could happen, just not what it took to clear it. Soooooo, when i powered up the 700 for the first time with your new board and virgin caps, that was the worst case scenario and the worst case happened. Drove me nuts there for a bit because it went away after trying it a couple more times, as i noted in my post when it happened.
Re: Jerry's " ONLY ONE IN THE COUNTRY" Phase Linear 700 Thre
No more surprises tonight Lee. I have some other work I have to get done. Will be back to experimenting with low end improvements in several days time.
PS: I especially like the sentence in the service manual that says:
"The model 700 has very little that can be done about turn-off noise aside from experimental matching of the front end differential transistors." We'll see...
No more surprises tonight Lee. I have some other work I have to get done. Will be back to experimenting with low end improvements in several days time.
PS: I especially like the sentence in the service manual that says:
"The model 700 has very little that can be done about turn-off noise aside from experimental matching of the front end differential transistors." We'll see...
- Joined
- Jan 14, 2011
- Messages
- 74,943
- Location
- Gillette, Wyo.
- Tagline
- Halfbiass...Electron Herder and Backass Woof
Re: Jerry's " ONLY ONE IN THE COUNTRY" Phase Linear 700 Thre
I've had some success matching but never cured it. I'm happy with the low end response of this amp anyway, HURTS ME to say it, but, this 700 also blows my Spec 2's out of the water. pain......pain........pain.......OH!!!!!! I have 2 700B's.....JOY.......JOY!.......JOY!!!!!!!!!!!!!!!!!!!!!
I've had some success matching but never cured it. I'm happy with the low end response of this amp anyway, HURTS ME to say it, but, this 700 also blows my Spec 2's out of the water. pain......pain........pain.......OH!!!!!! I have 2 700B's.....JOY.......JOY!.......JOY!!!!!!!!!!!!!!!!!!!!!
- Joined
- Jan 14, 2011
- Messages
- 74,943
- Location
- Gillette, Wyo.
- Tagline
- Halfbiass...Electron Herder and Backass Woof
Re: Jerry's " ONLY ONE IN THE COUNTRY" Phase Linear 700 Thre
Nando, that's just it, they have a very clean, tight bass, just not at 500 watts!!!!!!!!!!!!!!!!!!
When I get close to the bulb breaking levels on the Spec's the current limiters start to kick, it's not my speaks complaining cause I can put in one of these White Oak 1000's and get much louder and still stay clean. It's like after your first piece of ass or dope., there's no going back.
Nando, that's just it, they have a very clean, tight bass, just not at 500 watts!!!!!!!!!!!!!!!!!!
When I get close to the bulb breaking levels on the Spec's the current limiters start to kick, it's not my speaks complaining cause I can put in one of these White Oak 1000's and get much louder and still stay clean. It's like after your first piece of ass or dope., there's no going back.
- Joined
- Jan 14, 2011
- Messages
- 74,943
- Location
- Gillette, Wyo.
- Tagline
- Halfbiass...Electron Herder and Backass Woof
Re: Jerry's " ONLY ONE IN THE COUNTRY" Phase Linear 700 Thre
I was able to lean towards the Spec before Joe's board on the THD & N levels as the Spec had a much cleaner sound, that is no longer the case, these new boards will allow the PL's to stand toe to toe with about anything. I'm fixin to spend 800.00 on a Bryston 4B just to prove it to myself, then the 4B will be for sale.
I was able to lean towards the Spec before Joe's board on the THD & N levels as the Spec had a much cleaner sound, that is no longer the case, these new boards will allow the PL's to stand toe to toe with about anything. I'm fixin to spend 800.00 on a Bryston 4B just to prove it to myself, then the 4B will be for sale.