thanks for all the real world education. I've learned a lot of applicable skills from you that actually transfer out of the audio field. Many thanks my stranger friend.
I remember first seeing this "back-to-back" cap arrangement in "The Audiophiles Project Sourcebook" but the author did NOT go into ANY explanation. This was a great analysis of that type of arrangement.
Great video and explanations, as always, thank you! Very informative with practical measurements on the oscilloscope. Thank you also for the advice on speakers grounding and protection relays configurations, very important to learn. I will use your lessons in a vintage class A/B amp I have that needs restoring.
The negative of a capacitor is connected to the most outer layer of the plates. Thus, shielding is best on the outside and connected to the virtual ground. Even some non-electrolytic capacitors indicates the outer layer of the plates by means of a stripe. Tony, what you have learned over the years are absolutely correct. Bipolar capacitors are exactly as you explain. Connecting the positive leads together adds the equivalent inductances in series between the connection.
What is the final conclusion concerning the bipolar capacitor design? Are two polarized capacitors connected back to back equivalent to one bipolar capacitor? I say: No
Wow thanks so much for all the explanations, for those of us that are less familiar but want to do mods.. it's nice to know the why and how , not just "do this" . Great !
nice video update, thanks for sharing your knowledge. ps: 4 caps in a 2P2S configuration averages the ESR value between the caps so that the total resistance will be close the the ESR of just one cap.
.great video, thanks. Can you use 2 paralel bridge recifiers on the same transformer winding when thecircuits following the rectifier are completely isolated from each other, including the grounds not being shared?
Thanks for showing 2 electrolytic capacitors to form a non polarized capacitor. According to your scope trace, each independent polarized capacitor is still reverse biased. Is the scope setting on AC coupled or DC coupled? Thanks
Agreed - the scope appears to indicate both caps as being reversed biased - 'each for half of the AC cycle'. As for coupling, the scope display indicates that yellow (chan 1) and blue (chan 2) are indeed selected as DC coupled - as depicted with the symbols appearing just after the channel designation - you can also see chan 3 and chan 4 disabled for display and note the AC symbol for them.
@@TerranVisitor Thanks for the verification. If the polarized capacitors connected in series are still reverse biased, just wondering what is the difference of using one higher voltage polarized capacitor versus 2 lower voltage polarized capacitors connected in series? Say for example, a single 100v 100uf polarized cap. versus two 50v 200uf connected in series. Also when the applied frequency reduces (say less than 1Hz or even DC), will the polarized capacitors blow up under reverse bias?
The ones he is trying to explain are called balanced amplifier output where one output is half of the sine curve and the other the other part of the sine curve. No common gnd. between right and left. In order for them to detect 1.5v and higher, they are made so you have to have +- from the amplifier output. volt gnd- are also the same as speaker gnd, so it's not that easy to modify them. The delay they have is also nice to have to avoid bangs in the speaker at start-up
@@jim9930 Arhh, I understand now, It sounded like he was saying you shouldn't parallel 2 bridge rectifiers up, but if off a centre tapped transformer I can see there being a potential difference between each side, though I've never tried it.
Two things are of concern. First, is that you could have a different potential at ground, depending on the center tap of the coil. Also, diodes each have a different forward voltage drop, even though this could be a very small amount. This is still enough to cause one diode to conduct before the other. In addition, if you get the AC input leads of the two bridges reversed, you could potentially cause the Rectified AC (know known as ripple) to be 180 degrees out of phase, causing a dead short. If you care to try it, I would suggest putting a very low wattage dim bulb tester in line with the mains ;)
XrayTonyb What the Maximum Dc this board can handle? So both of my Amps I have are +- 100vdc and 80Vdc rails. Do I Need a Bigger voltage for 100uf BP capacitors which I need to order?
Tony's a smart man he keeps a spare motor start cap for his Air Conditioner, so If and when the condenser motor stops he can test the cap and if bad fix it himself cheap insurence / HVAC CALL SAY 200.00.
Say I have an amplifier or preamplifier that needs a 2x volt, balanced current system. The way I usually do it is to take + - so I don't load only one volt side and create an imbalance to the amplifier or preamplifier. Then I run them into a volt regulator with a larger heatsink and then into a card remember that the linear current regulator can only handle 1A and max 35v-40v linear volt regulator gives pure noise free current but gets hotter than other types.
I don't think it matters which way the capacitors connected up, the only benefit of connecting positive to positive is that you have the foil on the outside for EMI noise rejection what could be important. you can use the mains transformer for your little pcb if you tap it before the rectifier, form the lamp socket you're not interfering with the main secondary winding.
12:00 Well this shows just what was commented on in the first video, at the positive terminal of the capacitor (where the scope probes is connected), the voltage is more negative than the negative terminal of the capacitor half of the waveform. This is just a spreading the voltage over the 2 caps, not doing anything with the polarity of it. Try with two diodes, and se what the waveforms looks like then.
No need for AC or waveforms to complicate this. Just put a DC voltage across the so called ”bipolar” capacitor. Add a 1 M resistor across each cap to make sure the voltage is evenly distributed across the two capacitors, to not depend on unequal internal leakages. Measure the voltages across each capacitor with a voltmeter. One of the capacitors will be reverse biased, and will severely dislike the situation. I.e. This is not a bipolar capacitor.
@@andersjohansson8588 agreed. in his scope plot, the two probes provided a 10 Meg divider circuit. The display clearly shows reverse bias with 50% duty cycle on each cap... However, I have learned that most polarized electrolytic parts will behave fine and not suffer ill effects when the reverse bias voltage is
You're the "x-ray guy" :-) You could X-Ray one of those AC capacitors to show if they have back to back single capacitors or one large one. Just goofing with you but I worked in a Failure Analysis lab with real time X-Ray (Feinfocus/Cheetah) and we could do that and get a good image of what's inside. It's fun to play with the instrument.
27kΩ and 47uF make for a 6dB cut at 0.1Hz. That's enough to filter out even the largest bass signals so that they wouldn't reach the diode, get rectified and falsely trigger the protection circuit. Low ESR 47uF bipolar Nichicons are cheap enough so that you don't have to use two polar ones in series. Higher ESR will both increase the reaction time and also false trigger the protection circuit. Going cheap isn't worth it in this case. I would even bypass the bipolar with a 100nF ceramic, because the electrolytic ones tend to have higher ESR towards the supersonic frequencies and could also falsely trigger.
yes, it seems that with such a strong RC filter, the caps would only be reverse bias if the amplifier created a DC bias (or asymmetric clipping or there's very low freq from modulation side bands, etc). But isn't the point to detect problems like DC and not cause another failure?
Correct. As I said, you can adjust the values of the resistor or capacitor to set the cutoff wherever you want. In the case of the 27k/47uF, it would trip at DC, or a very high amount of hard clipping.
Correct. The standard test frequency for ESR is 100kHz. I had the DER set to 120 Hz, as I was connecting the caps to a 60Hz mains transformer for all the experiments.
@@xraytonyb At 120Hz it's not pure ESR i.e. the higher reactance lumps DF with ESR. But it's also indicative for the intended use. Still a valid comparison between caps if the test frequency is kept steady.
Regarding capacitors and AC. I always have difficulties when they say that capacitors block DC and pass AC current. SPOILERALERT !!! AC is also blocked by a capacitor! The current draw that you get between a AC source and a bipolair capacitor is not due the AC current going through the capacitor, but is the chargecurrent and discharge current of the capacitor. When u have a discharged electrolytic capacitor, and you connect it to a DC source, you get a surge current untill the capacitor is charged at same voltage potential as the DC source. When you flip the + and - of this DC source, and you connect it to the charged capacitor,got get a big spark, then the DC source will first discharge the capacitor to 0 volt, and then charge it back to the same potential as the DC source, but in this case with reversed polarity. When you keep flipping the polarity of this DC source, you keep getting charge and discharge current. This phenom of charging and discharging the DC capacitor is what happens with a bipolar capacitor and a AC source, the difference is that this bipolar capacitor is designed for this reversed polarity swap. A capacitor for DC or AC, is the same as 2 rooms (plates) separated by a wall (dielectric) but you can't go straight from one room to the other, and viceversa. The only way to go from one room to the other room, is by leaving one room and go through the hallway to the other room and vice versa. And as lost point, even a bipolar capacitor has a polarity. If you connect a bipolar capacitor like a 5uF , in parallel to the mains voltage, and you disconnect the mains voltage, regarding on which point of the 50/60Hz sine wave you disconnect the mains voltage, the bipolar capacitor will give out a DC voltage. When the disconnect happens at the top of the sine wave, the DC voltage will be near mains voltage. But when you discharge the bipolar capacitor, and flip the Phase and Neutral lead of the mains to the bipolar capacitor, and you disconnect the mains again, you will see that the DC polarity is still the same as before. I have done this experiment with different types of bipolar capacitors, and each time the polarity was the same, no mather how the Phase and Neutral where connected.
It's all a matter of semantics. In reality, a capacitor doesn't block or pass anything. It is a reactive device. Current leads the voltage (passes the signal, in a way) until the voltage of the capacitor equals the voltage source applied, at which point current no longer flows (blocks the signal, in a way). As for bipolar caps, a bipolar electrolytic is simply an electrolytic cap with two sets of electrodes configured in reverse series polarity and rolled up on top of one another. When voltage is applied to the bipolar cap in either direction, there will be an offset at the center of the four electrodes the will equal approximately 1/2 of the total voltage. With reference to the center, both caps will see a voltage in the proper polarity. Reversing the voltage source's polarity will cause the voltage across the cap to discharge/charge, leaving yet again, an offset voltage at the center. This is what was shown on the scope, although it may not at first seem as that is what we were seeing. Thanks for the comment!
@@xraytonyb the scope plots clearly show each capacitor with a negative voltage (reverse bias) for 50% of the time... large enough negative excursions will eventually cause degradation and failure
@@HazeAnderson , a frequency depended voltage divider is used in scope probes and in the frontend of a Differential scope probe to get a more flat band frequency response. Grtz
Hello, kind teacher. Please answer my question if you have a chance. Of course, I know that maybe this is not the place for this question, please forgive me I have a four-channel car amplifier, the driver part of the power transistors of one of the channels gets extremely hot, and the voltage on the base pin is about 22 volts. The other channel in the vicinity of this channel is the base pin voltage of power transistors on -2.5 V. If you answer that, I will be very happy.
In an amp without common ground on the outputs will not work. The dc reference for the outputs is done relatively to the ground. So, if you don't have a common ground, where is the reference gonna be? And, I think, an amp like Carver will have outputs protection. Maybe I'm wrong, who knows? 😁
The problem is that the polarity is reversed between the two channels and the phase is inverted on one channel to keep both channels in phase. This allows the amp to be wired for "bridged mono". If you try to tie the minus leads of the two channel together, you will short them out, as they will have opposite voltage polarity on them. You need to have isolated contacts for BOTH leads of the speakers with these amps.
@@xraytonyb, yes, understood, you must use an individual power supply, with separate ground for each module. Thank you for your time, I always enjoy watching your videos, always something to learn. And I appreciate the soldering talks.
![[LIVE] : ONE ลุมพินี 85 | คู่เอก "ยอดเหล็กเพชร vs ผึ้งหลวง"](http://i.ytimg.com/vi/Qq_19bnnxEo/mqdefault.jpg)
This is really great stuff!, This is the kind of stuff I ponder about sometimes. Thank you Tony!
thanks for all the real world education. I've learned a lot of applicable skills from you that actually transfer out of the audio field. Many thanks my stranger friend.
I remember first seeing this "back-to-back" cap arrangement in "The Audiophiles Project Sourcebook" but the author did NOT go into ANY explanation. This was a great analysis of that type of arrangement.
Great video and explanations, as always, thank you! Very informative with practical measurements on the oscilloscope. Thank you also for the advice on speakers grounding and protection relays configurations, very important to learn. I will use your lessons in a vintage class A/B amp I have that needs restoring.
The negative of a capacitor is connected to the most outer layer of the plates. Thus, shielding is best on the outside and connected to the virtual ground. Even some non-electrolytic capacitors indicates the outer layer of the plates by means of a stripe. Tony, what you have learned over the years are absolutely correct. Bipolar capacitors are exactly as you explain. Connecting the positive leads together adds the equivalent inductances in series between the connection.
What is the final conclusion concerning the bipolar capacitor design?
Are two polarized capacitors connected back to back equivalent to one bipolar capacitor?
I say: No
Very intersecting video. Thank you very much. Greetings' from Greece.
Wow thanks so much for all the explanations, for those of us that are less familiar but want to do mods.. it's nice to know the why and how , not just "do this" . Great !
nice video update, thanks for sharing your knowledge.
ps: 4 caps in a 2P2S configuration averages the ESR value between the caps so that the total resistance will be close the the ESR of just one cap.
.great video, thanks. Can you use 2 paralel bridge recifiers on the same transformer winding when thecircuits following the rectifier are completely isolated from each other, including the grounds not being shared?
Thanks for your life experiences & explanations
Nice presentation, Tony.
Thanks for showing 2 electrolytic capacitors to form a non polarized capacitor. According to your scope trace, each independent polarized capacitor is still reverse biased. Is the scope setting on AC coupled or DC coupled? Thanks
Agreed - the scope appears to indicate both caps as being reversed biased - 'each for half of the AC cycle'. As for coupling, the scope display indicates that yellow (chan 1) and blue (chan 2) are indeed selected as DC coupled - as depicted with the symbols appearing just after the channel designation - you can also see chan 3 and chan 4 disabled for display and note the AC symbol for them.
@@TerranVisitor Thanks for the verification. If the polarized capacitors connected in series are still reverse biased, just wondering what is the difference of using one higher voltage polarized capacitor versus 2 lower voltage polarized capacitors connected in series? Say for example, a single 100v 100uf polarized cap. versus two 50v 200uf connected in series. Also when the applied frequency reduces (say less than 1Hz or even DC), will the polarized capacitors blow up under reverse bias?
Thanks Tony!
The ones he is trying to explain are called balanced amplifier output where one output is half of the sine curve and the other the other part of the sine curve. No common gnd. between right and left.
In order for them to detect 1.5v and higher, they are made so you have to have +- from the amplifier output. volt gnd- are also the same as speaker gnd, so it's not that easy to modify them.
The delay they have is also nice to have to avoid bangs in the speaker at start-up
Nice and beautiful demonstrate sir👍
So, What happens when you put 2 bridge rectifiers in parallel on a transformer winding?
In my opinion, absolutely nothing. But I'd like to hear Tony's opinion. Just because he's got more experience and understanding than me.
@@jim9930 Arhh, I understand now, It sounded like he was saying you shouldn't parallel 2 bridge rectifiers up, but if off a centre tapped transformer I can see there being a potential difference between each side, though I've never tried it.
Two things are of concern. First, is that you could have a different potential at ground, depending on the center tap of the coil. Also, diodes each have a different forward voltage drop, even though this could be a very small amount. This is still enough to cause one diode to conduct before the other. In addition, if you get the AC input leads of the two bridges reversed, you could potentially cause the Rectified AC (know known as ripple) to be 180 degrees out of phase, causing a dead short. If you care to try it, I would suggest putting a very low wattage dim bulb tester in line with the mains ;)
XrayTonyb What the Maximum Dc this board can handle? So both of my Amps I have are +- 100vdc and 80Vdc rails. Do I Need a Bigger voltage for 100uf BP capacitors which I need to order?
Excellent.
You are the master !
Tony's a smart man he keeps a spare motor start cap for his Air Conditioner, so If and when the condenser motor stops he can test the cap and if bad fix it himself cheap insurence / HVAC CALL SAY 200.00.
Buying a spare capacitor before it fails is called being smart. Buying one after it has failed once before is called being experienced ;)
@@xraytonyb amen! Enjoyed watching.
Say I have an amplifier or preamplifier that needs a 2x volt, balanced current system.
The way I usually do it is to take + - so I don't load only one volt side and create an imbalance to the amplifier or preamplifier.
Then I run them into a volt regulator with a larger heatsink and then into a card
remember that the linear current regulator can only handle 1A and max 35v-40v
linear volt regulator gives pure noise free current but gets hotter than other types.
I don't think it matters which way the capacitors connected up, the only benefit of connecting positive to positive is that you have the foil on the outside for EMI noise rejection what could be important.
you can use the mains transformer for your little pcb if you tap it before the rectifier, form the lamp socket you're not interfering with the main secondary winding.
12:00 Well this shows just what was commented on in the first video,
at the positive terminal of the capacitor (where the scope probes is connected),
the voltage is more negative than the negative terminal of the capacitor half of the waveform.
This is just a spreading the voltage over the 2 caps, not doing anything with the polarity of it.
Try with two diodes, and se what the waveforms looks like then.
Yes - at 10:13 he clearly shows each capacitor with a negative voltage for 50% of the time
No need for AC or waveforms to complicate this.
Just put a DC voltage across the so called ”bipolar” capacitor.
Add a 1 M resistor across each cap to make sure the voltage is
evenly distributed across the two capacitors, to not depend on
unequal internal leakages.
Measure the voltages across each capacitor with a voltmeter.
One of the capacitors will be reverse biased, and will severely dislike the situation.
I.e. This is not a bipolar capacitor.
@@andersjohansson8588 agreed. in his scope plot, the two probes provided a 10 Meg divider circuit. The display clearly shows reverse bias with 50% duty cycle on each cap...
However, I have learned that most polarized electrolytic parts will behave fine and not suffer ill effects when the reverse bias voltage is
Xraytonyb your utube videos are awesome
You're the "x-ray guy" :-) You could X-Ray one of those AC capacitors to show if they have back to back single capacitors or one large one. Just goofing with you but I worked in a Failure Analysis lab with real time X-Ray (Feinfocus/Cheetah) and we could do that and get a good image of what's inside. It's fun to play with the instrument.
27kΩ and 47uF make for a 6dB cut at 0.1Hz. That's enough to filter out even the largest bass signals so that they wouldn't reach the diode, get rectified and falsely trigger the protection circuit. Low ESR 47uF bipolar Nichicons are cheap enough so that you don't have to use two polar ones in series. Higher ESR will both increase the reaction time and also false trigger the protection circuit. Going cheap isn't worth it in this case. I would even bypass the bipolar with a 100nF ceramic, because the electrolytic ones tend to have higher ESR towards the supersonic frequencies and could also falsely trigger.
yes, it seems that with such a strong RC filter, the caps would only be reverse bias if the amplifier created a DC bias (or asymmetric clipping or there's very low freq from modulation side bands, etc). But isn't the point to detect problems like DC and not cause another failure?
Correct. As I said, you can adjust the values of the resistor or capacitor to set the cutoff wherever you want. In the case of the 27k/47uF, it would trip at DC, or a very high amount of hard clipping.
Xraytonyb your information is intresting about speaker protect circuit
Xraytonyb you are good at electronics
ESR is usually specified at 100kHz? You may also use the DER EE's Rs mode for that.
Correct. The standard test frequency for ESR is 100kHz. I had the DER set to 120 Hz, as I was connecting the caps to a 60Hz mains transformer for all the experiments.
@@xraytonyb At 120Hz it's not pure ESR i.e. the higher reactance lumps DF with ESR. But it's also indicative for the intended use. Still a valid comparison between caps if the test frequency is kept steady.
Regarding capacitors and AC. I always have difficulties when they say that capacitors block DC and pass AC current.
SPOILERALERT !!! AC is also blocked by a capacitor!
The current draw that you get between a AC source and a bipolair capacitor is not due the AC current going through the capacitor,
but is the chargecurrent and discharge current of the capacitor.
When u have a discharged electrolytic capacitor, and you connect it to a DC source, you get a surge current untill the capacitor is charged at same voltage potential as the DC source. When you flip the + and - of this DC source, and you connect it to the charged capacitor,got get a big spark, then the DC source will first discharge the capacitor to 0 volt, and then charge it back to the same potential as the DC source, but in this case with reversed polarity. When you keep flipping the polarity of this DC source, you keep getting charge and discharge current.
This phenom of charging and discharging the DC capacitor is what happens with a bipolar capacitor and a AC source, the difference is that this bipolar capacitor is designed for this reversed polarity swap. A capacitor for DC or AC, is the same as 2 rooms (plates) separated by a wall (dielectric) but you can't go straight from one room to the other, and viceversa. The only way to go from one room to the other room, is by leaving one room and go through the hallway to the other room and vice versa.
And as lost point, even a bipolar capacitor has a polarity. If you connect a bipolar capacitor like a 5uF , in parallel to the mains voltage, and you disconnect the mains voltage, regarding on which point of the 50/60Hz sine wave you disconnect the mains voltage, the bipolar capacitor will give out a DC voltage. When the disconnect happens at the top of the sine wave, the DC voltage will be near mains voltage. But when you discharge the bipolar capacitor, and flip the Phase and Neutral lead of the mains to the bipolar capacitor, and you disconnect the mains again, you will see that the DC polarity is still the same as before.
I have done this experiment with different types of bipolar capacitors, and each time the polarity was the same, no mather how the Phase and Neutral where connected.
"AC is also blocked by a capacitor!" Negligible but add a resistor and you get a frequency dependent voltage divider ... AKA a filter.
It's all a matter of semantics. In reality, a capacitor doesn't block or pass anything. It is a reactive device. Current leads the voltage (passes the signal, in a way) until the voltage of the capacitor equals the voltage source applied, at which point current no longer flows (blocks the signal, in a way). As for bipolar caps, a bipolar electrolytic is simply an electrolytic cap with two sets of electrodes configured in reverse series polarity and rolled up on top of one another. When voltage is applied to the bipolar cap in either direction, there will be an offset at the center of the four electrodes the will equal approximately 1/2 of the total voltage. With reference to the center, both caps will see a voltage in the proper polarity. Reversing the voltage source's polarity will cause the voltage across the cap to discharge/charge, leaving yet again, an offset voltage at the center. This is what was shown on the scope, although it may not at first seem as that is what we were seeing. Thanks for the comment!
@@xraytonyb the scope plots clearly show each capacitor with a negative voltage (reverse bias) for 50% of the time... large enough negative excursions will eventually cause degradation and failure
@@HazeAnderson , a frequency depended voltage divider is used in scope probes and in the frontend of a Differential scope probe to get a more flat band frequency response. Grtz
@@xraytonyb thank you for the reply. Grtz
👍👍 super
Was that a piece of Cheese-It on the paper at the beginning? Sweet!
Not a food item (although that hap[pens sometimes). Most likely a piece of plastic or glue from a PC board.
Hello, kind teacher. Please answer my question if you have a chance. Of course, I know that maybe this is not the place for this question, please forgive me
I have a four-channel car amplifier, the driver part of the power transistors of one of the channels gets extremely hot, and the voltage on the base pin is about 22 volts.
The other channel in the vicinity of this channel is the base pin voltage of power transistors on -2.5 V.
If you answer that, I will be very happy.
In an amp without common ground on the outputs will not work. The dc reference for the outputs is done relatively to the ground. So, if you don't have a common ground, where is the reference gonna be? And, I think, an amp like Carver will have outputs protection. Maybe I'm wrong, who knows? 😁
@@jim9930, I red your comment. We were talking about those protect circuits. Fuse protection woks fine on hard shorts, but not on small DC offset.
The problem is that the polarity is reversed between the two channels and the phase is inverted on one channel to keep both channels in phase. This allows the amp to be wired for "bridged mono". If you try to tie the minus leads of the two channel together, you will short them out, as they will have opposite voltage polarity on them. You need to have isolated contacts for BOTH leads of the speakers with these amps.
@@xraytonyb, yes, understood, you must use an individual power supply, with separate ground for each module. Thank you for your time, I always enjoy watching your videos, always something to learn. And I appreciate the soldering talks.
Great info, thanks.