Reading post 52 again it seems you cannot "center" the output to ground with R15
and the given voltages. It may be that they set the values for maximum output
excursion capability. This can be checked with input signal, output load and scope,
ungrounded or with differential input. Symmetrical clipping would be best.
But a deviation of a few volts as indicated by you is too much.
In post 52, after HOWEVER read output instead of input ?
and the given voltages. It may be that they set the values for maximum output
excursion capability. This can be checked with input signal, output load and scope,
ungrounded or with differential input. Symmetrical clipping would be best.
But a deviation of a few volts as indicated by you is too much.
In post 52, after HOWEVER read output instead of input ?
I have new zeners arriving later today so I'll swap them and we'll see what kind of result we get. Fairly certain the zeners in series that should give me -58V are both open so the voltage drop I am getting is simply because of the 150ohm resistor. I had enough foresight to order both 51 and 58 volt zeners so I'll try returning to the schematic first and if I have any issue I'll go to the 58v zener.
Your understanding is correct about what I'm trying to accomplish and thank you very much for your help. Hopefully I can get this sorted tonight.
Once I replaced the two in series zeners with a 56V Zener (I ordered a 56v and not a 58v) I was able to get the DC balance pretty close to zero by maxing out the 25K trimmer pot (within .5V). What I have noticed is that the amplifier as it warms up tends to consume less power so my guess is that the DC balance needs to be set once the amplifier is at operating temperature or it'll just drift away as it warms up, so I'll need to let it warm up and see if the 56V zener is the right choice or if I'll need to go to the 51V zener. My guess is that I'll up swapping the 56v for the 51v since I think less power consumption means the DC balance is moving away from the direction I need it to.
Last thing I did today after fixing the power supply was replace the phase splitter for each side of the amplifier. I used a beta matched pair of JAN2N2219. I did this since one of the phase splitter transistors had already been replaced with a harsh sounding 2N3241. Once that replacement was made my sound and output are perfectly balanced.
This amplifier is great! By comparison, a Marantz 15 sounds like a noisy harsh mess. Nothing about this amplifier is fatiguing and I could sit here and listen for hours without a problem. There's a resolution here that is just unmatched. I can hear mistakes and background noises in recordings that I don't think I've ever heard before. This is a great design which is why I chose to put so much time into my two copies of this amplifier and I'm glad I did. It sounds like absolutely nothing.
Last thing I did today after fixing the power supply was replace the phase splitter for each side of the amplifier. I used a beta matched pair of JAN2N2219. I did this since one of the phase splitter transistors had already been replaced with a harsh sounding 2N3241. Once that replacement was made my sound and output are perfectly balanced.
This amplifier is great! By comparison, a Marantz 15 sounds like a noisy harsh mess. Nothing about this amplifier is fatiguing and I could sit here and listen for hours without a problem. There's a resolution here that is just unmatched. I can hear mistakes and background noises in recordings that I don't think I've ever heard before. This is a great design which is why I chose to put so much time into my two copies of this amplifier and I'm glad I did. It sounds like absolutely nothing.
Last edited:
Hi hadley,
Well, the 2SC3421 isn`t a harsh sounding transistor. But if the characteristics don`t match the circuit you could get higher distortion. Also, it may not have been a real 2SC3421. Blanket statements like that don`t have any basis in fact. If a transistor is good quality, it doesn`t have a `sound`. It`s the interaction with the circuit design and values do.
When you attempt to match transistors, they must be held at exactly the same temperature or you simply cannot match them. If you look at one of the Adcom GFA-565 threads, you may find a matcher design I gave to the community. A few board designs have been made for it and it has been used to great success by many members already. Anyway, the critical point about differential pairs is that they should be maintained at the exact same temperature or as close as you can get, and the same is true when you attempt to match them.
-Chris
Well, the 2SC3421 isn`t a harsh sounding transistor. But if the characteristics don`t match the circuit you could get higher distortion. Also, it may not have been a real 2SC3421. Blanket statements like that don`t have any basis in fact. If a transistor is good quality, it doesn`t have a `sound`. It`s the interaction with the circuit design and values do.
When you attempt to match transistors, they must be held at exactly the same temperature or you simply cannot match them. If you look at one of the Adcom GFA-565 threads, you may find a matcher design I gave to the community. A few board designs have been made for it and it has been used to great success by many members already. Anyway, the critical point about differential pairs is that they should be maintained at the exact same temperature or as close as you can get, and the same is true when you attempt to match them.
-Chris
-I didn't have any difficulty matching any transistors.
-This circuit isn't even sensitive to beta outside of the phase splitter.
-This 622C doesn't have a 2SC3421, the 2N3421 was not original and the beta was nowhere near where it needed to be.
-The Hadley has the differential pair housed in a single TO-78 MAT12AHZ Dual NPN, so by definition the temperature for both would be the same.
Given these points, I have zero idea what you're talking about.
Last edited:
Hi hadley,
Your matches are probably not that good, but good enough for you. Close matched transistors will lower distortion and do sound better. For you, as long as it functions, you're happy.
Take a little time and research this a little. You will see what I mean. I am trying to assist and educate you. Several other members have looked into it and understand perfectly now.
Anyway, gave you some true facts. I can't do anything more.
as_audio is exactly right about the phase splitter specifically. This generally holds true everywhere. You have to consider capacitances, high frequency cut-off points, beta and beta linearity. But sound quality is generally determined by PCB layout, circuit design and some components. Circuits are designed so that transistor characteristics have reduced impact on performance. Same for tube circuits.
-Chris
Exactly my point.
Your matches are probably not that good, but good enough for you. Close matched transistors will lower distortion and do sound better. For you, as long as it functions, you're happy.
Take a little time and research this a little. You will see what I mean. I am trying to assist and educate you. Several other members have looked into it and understand perfectly now.
Anyway, gave you some true facts. I can't do anything more.
as_audio is exactly right about the phase splitter specifically. This generally holds true everywhere. You have to consider capacitances, high frequency cut-off points, beta and beta linearity. But sound quality is generally determined by PCB layout, circuit design and some components. Circuits are designed so that transistor characteristics have reduced impact on performance. Same for tube circuits.
-Chris
Yup, sorry ...
Very often a production amplifier does not have matched pairs in the diff pair. position You can really improve performance by installing a matched pair. Thinking back, this amp used a monolithic pair, didn't it? He did replace it with a modern, improved part.
The only reason to use a matched pair is in the diff pair - which this amp has. I assumed he was talking about that. Matching the phase splitter transistors makes little to no sense so I assumed he was talking about the one place it matters.
Very often a production amplifier does not have matched pairs in the diff pair. position You can really improve performance by installing a matched pair. Thinking back, this amp used a monolithic pair, didn't it? He did replace it with a modern, improved part.
Does this at all matter when we take into account the rest of the amplifier? Not likely. But, this is the right way to do it as far as I'm aware.
Last edited:
Hi as_audio,
I indicated I erred in making an assumption and explained it fully. So I was not trying to confuse things.
Most of us are doing much more than focusing on a single issue or thread, and this is the case with me. I'm designing or repairing other amplifiers as my main activity. So the specific details of this thread or others are not foremost in my mind all the time. I make an honest effort to both assist people, and keeping the information technically correct.
In that case I did admit to making a mistake out in the open, so pursuing this further is something I do not understand. I did state that I recalled he replaced the dual transistors - remember?
So let's be clear. I made an incorrect assumption because the OP mentioned using a matched pair. Something not needed in that circuit position.
Is that what you wanted to hear? Are we good now?
-Chris
I indicated I erred in making an assumption and explained it fully. So I was not trying to confuse things.
Most of us are doing much more than focusing on a single issue or thread, and this is the case with me. I'm designing or repairing other amplifiers as my main activity. So the specific details of this thread or others are not foremost in my mind all the time. I make an honest effort to both assist people, and keeping the information technically correct.
In that case I did admit to making a mistake out in the open, so pursuing this further is something I do not understand. I did state that I recalled he replaced the dual transistors - remember?
So let's be clear. I made an incorrect assumption because the OP mentioned using a matched pair. Something not needed in that circuit position.
Is that what you wanted to hear? Are we good now?
-Chris
The Marantz 15 was actually one of the last made in NY. It was developed before the company was sold to Superscope.
I have a lot of tube gear and have rebuilt tube amplifiers for a few years, but both of these 622s are my first attempts at completely rebuilding solid state equipment. I thought if I could get the rougher of the two 622s working that I'd know enough by then to do the second, which is what I've done. Other amplifiers I have include a Marantz 8b, the McIntosh MC225 and while both are amazing amplifiers in their own way, they just do not have the low-end authority this solid state amplifier does. I'll have to pull them out and do some more testing before I can say more.
Once I replaced the two in series zeners with a 56V Zener (I ordered a 56v and not a 58v) I was able to get the DC balance pretty close to zero by maxing out the 25K trimmer pot (within .5V). What I have noticed is that the amplifier as it warms up tends to consume less power so my guess is that the DC balance needs to be set once the amplifier is at operating temperature or it'll just drift away as it warms up, so I'll need to let it warm up and see if the 56V zener is the right choice or if I'll need to go to the 51V zener. My guess is that I'll up swapping the 56v for the 51v since I think less power consumption means the DC balance is moving away from the direction I need it to.
Last thing I did today after fixing the power supply was replace the phase splitter for each side of the amplifier. I used a beta matched pair of JAN2N2219. I did this since one of the phase splitter transistors had already been replaced with a harsh sounding 2N3241. Once that replacement was made my sound and output are perfectly balanced.
This amplifier is great! By comparison, a Marantz 15 sounds like a noisy harsh mess. Nothing about this amplifier is fatiguing and I could sit here and listen for hours without a problem. There's a resolution here that is just unmatched. I can hear mistakes and background noises in recordings that I don't think I've ever heard before. This is a great design which is why I chose to put so much time into my two copies of this amplifier and I'm glad I did. It sounds like absolutely nothing.
Finally accumulated enough in my Mouser basket to finish my order containing the 1.3W 51V zener diodes. Since my last post, the amplifier dropped a channel. Thankfully, when I opened up the amplifier I was able to determine it was just a cold solder joint on one of the 8uf capacitors. While I was in there, I swapped out the 0.5W 6.8V and 56V Zeners for the 1.3W 6.8V and 51V the schematic called for as we determined earlier. Fiddled with the amplifier for a while to get everything set just right and ended up with centered 25K and 200 ohm pots. With the dual NPN differential pair and all of the matched parts, the 200 ohm pot is unnecessary and could be removed. As before, the amplifier sounds excellent and now I hope it is reliable.
All that's left to do now is finish up the other, earlier 622C amplifier I started working on first and this Hadley project will be done.
Had a look at the schematic again (post 1).
This amplifier has no conventional second / "driver" stage at all,
the input diff stage does all the voltage swing.
It would be interesting to know, from measurement, the max output swing.
Easiest way to determine is with a triangle wave form and scope.
Schematic in post 1 seems to be on a different server - is it safe in the future ?
This amplifier has no conventional second / "driver" stage at all,
the input diff stage does all the voltage swing.
It would be interesting to know, from measurement, the max output swing.
Easiest way to determine is with a triangle wave form and scope.
Schematic in post 1 seems to be on a different server - is it safe in the future ?
I'll look at measuring this, but likely will not push the amplifier beyond a 1V input though. I think it is theoretically possible to push the MAT12s beyond their VCEO in this circuit and since they're $35/each I'd like not to damage them.
I am hosting the schematic on my own server and have copies of it saved.