rechecked my batch of transistors and sorted a pair of 5551's that were really close, popped them in and now am looking at under 10mv.
time to hook up meters for bias and dc and let it run for a while, reset the bias and then see where we are at.
time to hook up meters for bias and dc and let it run for a while, reset the bias and then see where we are at.
been running for over 45min now, bias stabilised at 290mv, the right channel is now floating around 1mv with left channel at 6-7mv still.
running on the bench with some old spkrs, will let it go for a while then cool it down and power up from cold to see where it's at.
running on the bench with some old spkrs, will let it go for a while then cool it down and power up from cold to see where it's at.
Today I decided to revisit my BG 470uF bipolar caps, figured couldnt be anything wrong with them, retested all good so tentatively fitted them in one channel at a time, testing each stage. 6mv/3mv DC, amp is now running in again with the black gates, they will take some time, anyone who has experience with the red bipolar range will know the sometimes strange sounds you get from them during the first hours of use 🙂
Hi Sonic Art,
Early transistor manufacturing processes were not nearly as well controlled as they are today. I would buy 20 pcs to match two pairs and sometimes may get a few more matches. Early high end service was hell (and expensive). Output transistors were also all over the map, same story to match them (yes, it makes a difference). Marantz and a few other manufacturers knew this and offered matched gain transistors, and they were clear about it in their service manuals. You have n idea how high the volume of semiconductors I bought in the early days! Thank goodness for the cheap stuff where it didn't matter as I could use up the ones that didn't have matches.
Break - in is nonsense, sorry a complete myth. What is happening is that you are getting use to the sound. The part is not changing audibly, and no you cannot hear those changes.
Can you imagine how industry would deal with break - in at the levels you can hear??? Test equipment, oscillators and voltage references do settle - but at tiny, high accuracy levels well below what you could ever hear. Resistors may age, and change value with impressed voltage and temperature. They always do, some more than others. Not enough for you to hear. Temperature is the biggest variable, aging occurs slowly, most in early hours. This is all an echo from the "I can hear things that can't be measured" group. This is not the case anymore, and hasn't been for a couple decades anyway.
I still buy signal transistors 100 at a time. I hate matching them, but it makes a big difference - so I do it.
-Chris
Early transistor manufacturing processes were not nearly as well controlled as they are today. I would buy 20 pcs to match two pairs and sometimes may get a few more matches. Early high end service was hell (and expensive). Output transistors were also all over the map, same story to match them (yes, it makes a difference). Marantz and a few other manufacturers knew this and offered matched gain transistors, and they were clear about it in their service manuals. You have n idea how high the volume of semiconductors I bought in the early days! Thank goodness for the cheap stuff where it didn't matter as I could use up the ones that didn't have matches.
Break - in is nonsense, sorry a complete myth. What is happening is that you are getting use to the sound. The part is not changing audibly, and no you cannot hear those changes.
Can you imagine how industry would deal with break - in at the levels you can hear??? Test equipment, oscillators and voltage references do settle - but at tiny, high accuracy levels well below what you could ever hear. Resistors may age, and change value with impressed voltage and temperature. They always do, some more than others. Not enough for you to hear. Temperature is the biggest variable, aging occurs slowly, most in early hours. This is all an echo from the "I can hear things that can't be measured" group. This is not the case anymore, and hasn't been for a couple decades anyway.
I still buy signal transistors 100 at a time. I hate matching them, but it makes a big difference - so I do it.
-Chris
it's OK, as a tech/engineer you are going to have your own bias (pardon the pun) but growth mindset is not reserved purely for our kids 😉
I appreciate any help given and your time, end of the day I figured it out and learned a lot despite being a "hacker" (as seems to be the term for those who chuck components at equipment in order to "upgrade" the sound) :d
many of my questions were not answered/overlooked, that's disappointing as I am sure there are many reading these threads and hoping for an answer they too can understand
for example, updating the drain resistors to 470/220ohm and fitting the 0.1uf caps to a DH200 with no other circuit changes, aka DH220, is this a good move? why?
I appreciate any help given and your time, end of the day I figured it out and learned a lot despite being a "hacker" (as seems to be the term for those who chuck components at equipment in order to "upgrade" the sound) :d
many of my questions were not answered/overlooked, that's disappointing as I am sure there are many reading these threads and hoping for an answer they too can understand
for example, updating the drain resistors to 470/220ohm and fitting the 0.1uf caps to a DH200 with no other circuit changes, aka DH220, is this a good move? why?
« Updating the drain resistors to 470/220ohm » is a good move to have more equal response time between N and P channels because of the different mosfet input capacitance (time constant —> R*C) , it is a known fact documented in many threads and technical papers. Although not mandatory since it works fine anyway. But in the end, the sum of all changes as an « improvement« has the potential to make a sound difference…..
I prefer to not discuss « breakin« of some type of components since it is a known controversial subject on forums….😉
not sure what 0.1uf cap you refer to….🤔
Fab
I prefer to not discuss « breakin« of some type of components since it is a known controversial subject on forums….😉
not sure what 0.1uf cap you refer to….🤔
Fab
Thanks Fab, and while it may be documented, and I have read 90% of this thread as well as every other I can find on Hafler, I did not see whether there was any need to change anything else in the circuit in order to implement this "update" from DH200 to DH220. I dont want to make changes unless I am sure they will not upset anything else or I understand what else is required to implement them.
0.1uf, sorry drain? capacitors, DH200 has a single 680pf mica from the N side only, 220 has the 0.1 both sides to ground.
And C13/C401, that was changed from 390pf to 680pf in the DH220. This cap goes from source to gate yes?
And the resistors are from gate (not drain) to ground right? need to keep my terminology correct so please let me know if I use the wrong words when trying to convey things! 😉 To be clear, MY resistors are all currently 470ohm, they look original but may have been changed during the amps life.
Oh and in my notes I also have that: "in the DH220 a 120pF cap was added from collector to base of Q13"
is this a worthwhile addition to Q13 in the DH200?
0.1uf, sorry drain? capacitors, DH200 has a single 680pf mica from the N side only, 220 has the 0.1 both sides to ground.
And C13/C401, that was changed from 390pf to 680pf in the DH220. This cap goes from source to gate yes?
And the resistors are from gate (not drain) to ground right? need to keep my terminology correct so please let me know if I use the wrong words when trying to convey things! 😉 To be clear, MY resistors are all currently 470ohm, they look original but may have been changed during the amps life.
Oh and in my notes I also have that: "in the DH220 a 120pF cap was added from collector to base of Q13"
is this a worthwhile addition to Q13 in the DH200?
Last edited:
Amp has been running nicely, but it seemed a little veiled, bugged me when I know from having restored someone elses DH220 what it should sound like.
So today I swapped out the P channel gate resistors to 220ohm, source to gate cap to 680pf, and added the 0.1uf from drain to ground. I used dayton PPT polyprop, not metallised poly.
I also pulled the vishay bypass off my 220uf caps & the Vishay C20/21 and replaced with wima MKP10 0.01uf 630v I had on hand.
End result much more clarity & transparency, tighter imaging, starting to sound like a tricked up Hafler now!
Still like to know about the cap across Q13 and if its worth ordering some 120pf silver mica to emulate the DH220 circuit?
Gate resistor quality, how important is this in the Hafler? should I look to fit higher quality metal film or are the carbon good enough?
So today I swapped out the P channel gate resistors to 220ohm, source to gate cap to 680pf, and added the 0.1uf from drain to ground. I used dayton PPT polyprop, not metallised poly.
I also pulled the vishay bypass off my 220uf caps & the Vishay C20/21 and replaced with wima MKP10 0.01uf 630v I had on hand.
End result much more clarity & transparency, tighter imaging, starting to sound like a tricked up Hafler now!
Still like to know about the cap across Q13 and if its worth ordering some 120pf silver mica to emulate the DH220 circuit?
Gate resistor quality, how important is this in the Hafler? should I look to fit higher quality metal film or are the carbon good enough?
Hi Sonic Art,
You only see current flow through gate stoppers at higher frequencies, there is no DC current flow.
So non inductive metal film resistors should be used here (I have seen inductive ones!). A carbon composition resistor should also be fine - but they can have their own problems that new manufactured ones should avoid.
The best capacitors are polystyrene and polypropylene. Any decent brand will be equivalent as long as they are not too large. Anything else are stories and hype, and you are welcome to pay extra for that.
-Chris
You only see current flow through gate stoppers at higher frequencies, there is no DC current flow.
So non inductive metal film resistors should be used here (I have seen inductive ones!). A carbon composition resistor should also be fine - but they can have their own problems that new manufactured ones should avoid.
The best capacitors are polystyrene and polypropylene. Any decent brand will be equivalent as long as they are not too large. Anything else are stories and hype, and you are welcome to pay extra for that.
-Chris
Chris: For op amp bypass at the pins of the device, I understand from Tom Christiansen's Neurochrome that NPO or COG ceramic capacitors are best due to their high frequency response. What is your view?
Also, when you refer to "large", I assume you are referring to physical size. Do you have some rule of thumb that relates capacity to max physical size, or is it a hard physical limit and what would your guidance be?
Hi Halauhula,
Yes, good NP0 / C0G ceramic capacitors are every bit as good. The problem is that I measure variable quality with ceramics even from authorized distributors. I do use ceramic capacitors - but I measure each one first (I do with all parts anyway). I don't see that issue with the part types I mentioned, however if you buy from sources that are not in the authorized distribution chain you can get iffy quality.
Yup. If a part is physically too large, it will end up close to things it shouldn't be. A capacitor is a big antenna and can transmit or receive. So it doesn't make any sense to use a huge capacitor. When the equipment was designed, they made allowances for where the capacitor mounted. Smaller is great and is normally the case these days. This is not true for electrolytic capacitors by the way. As long as the part is not larger than the original you should be fine.
I have even seen film capacitors installed where they actually restrict airflow. This is stupid beyond belief! Common in tube amp upgrades.
Yes, good NP0 / C0G ceramic capacitors are every bit as good. The problem is that I measure variable quality with ceramics even from authorized distributors. I do use ceramic capacitors - but I measure each one first (I do with all parts anyway). I don't see that issue with the part types I mentioned, however if you buy from sources that are not in the authorized distribution chain you can get iffy quality.
Yup. If a part is physically too large, it will end up close to things it shouldn't be. A capacitor is a big antenna and can transmit or receive. So it doesn't make any sense to use a huge capacitor. When the equipment was designed, they made allowances for where the capacitor mounted. Smaller is great and is normally the case these days. This is not true for electrolytic capacitors by the way. As long as the part is not larger than the original you should be fine.
I have even seen film capacitors installed where they actually restrict airflow. This is stupid beyond belief! Common in tube amp upgrades.
Thank you for your advice. As a hobbyist, I do not have access to dedicated cap checkers. May I assume that using capacitor check function (which only measures capacity) on a multimeter is OK? Or is another type of check recommended and how would I do that? Here, I am referring to bypass, DC blocking and EQ/filter capacitors, not PS capacitors.
Hi Halauhula,
Fair enough.
No. You need to measure things like D-A at various frequencies. That means a real, and expensive, LCR meter. I use a few, HP 4263A and a nice 4192A, also Genrad and a few others. Cheap instruments do not give you dependable answers - I have some, I have tried to find a cheap solution I can recommend.
Just buy good types of parts from authorized distributors. That is your best defense and there is a very good reason I do the same. Common sense is your best friend as a hobbyist. Never fall into the trap of thinking just because you are a hobbyist, cheap stuff is okay. The rules of physics don't change just because someone is a hobbyist, and going cheap is wasting money. None of us can afford that. If you do, at least recognise the limitations of what you did buy. However, that takes training and experience and we all want to be optimistic. By the way, the last digit on a digital meter is usually complete fiction. Very good meters give much closer answers, but become use to thinking the word "about" before reading the display. I have seen meters where the last two digits may as well have not been there - and that was if they were in tolerance! I also come from a lab test and measurement background and calibrated meters and things. So what I am telling you is observed fact.
-Chris
Fair enough.
No. You need to measure things like D-A at various frequencies. That means a real, and expensive, LCR meter. I use a few, HP 4263A and a nice 4192A, also Genrad and a few others. Cheap instruments do not give you dependable answers - I have some, I have tried to find a cheap solution I can recommend.
Just buy good types of parts from authorized distributors. That is your best defense and there is a very good reason I do the same. Common sense is your best friend as a hobbyist. Never fall into the trap of thinking just because you are a hobbyist, cheap stuff is okay. The rules of physics don't change just because someone is a hobbyist, and going cheap is wasting money. None of us can afford that. If you do, at least recognise the limitations of what you did buy. However, that takes training and experience and we all want to be optimistic. By the way, the last digit on a digital meter is usually complete fiction. Very good meters give much closer answers, but become use to thinking the word "about" before reading the display. I have seen meters where the last two digits may as well have not been there - and that was if they were in tolerance! I also come from a lab test and measurement background and calibrated meters and things. So what I am telling you is observed fact.
-Chris
I only buy from Mouser, Digikey, Newark, Michael Percy and Parts Express, for the very reason of getting only legitimate parts (I hope). Thank you for your advice. I will see what prices are for a used HP4263A or 4192A from a legitimate refurbisher. Any other models like the Gerard would be appreciated as I do a search.
ESI 2150 (I use one) or similar.
These are not cheap. Mind you, they should last a very, very long time. I have LCR bridges as well by Genrad and others. They still work great, but the newer digital ones are faster and easier to read without making errors if you're having a off day. You also need to understand how to use them and other factors. Plus what is normal for a part type. This will be an excellent education and will help in all aspects of electronics. Also, PCB material has losses it would be good to be aware of.
Don't forget to buy the test fixtures. My Kelvin leads for the HP 4263A cost me about $600 when I bought the meter new in the mid 80's ( I think). The 4263A was a brand new model then when I bought it. The 4192A test fixtures cost me around $1,000, and that was recent. You don't need them all, but be aware the leads and fixtures are important and directly impact your measurements. So cheap ones will degrade the instrument. You need at least the Kelvin lead set, and I use the axial lead fixture often also.
-Chris
These are not cheap. Mind you, they should last a very, very long time. I have LCR bridges as well by Genrad and others. They still work great, but the newer digital ones are faster and easier to read without making errors if you're having a off day. You also need to understand how to use them and other factors. Plus what is normal for a part type. This will be an excellent education and will help in all aspects of electronics. Also, PCB material has losses it would be good to be aware of.
Don't forget to buy the test fixtures. My Kelvin leads for the HP 4263A cost me about $600 when I bought the meter new in the mid 80's ( I think). The 4263A was a brand new model then when I bought it. The 4192A test fixtures cost me around $1,000, and that was recent. You don't need them all, but be aware the leads and fixtures are important and directly impact your measurements. So cheap ones will degrade the instrument. You need at least the Kelvin lead set, and I use the axial lead fixture often also.
-Chris
thanks Chris, the Dayton caps are nothing special, at least not on the audiophile radar, 2 bucks each will hardly break the bank 😉
they are not all that big either, fit with plenty of room around them in the heatsink gap.
I will pick up some 0.6w metal films, any particular brand you prefer for its specs and reliability?
I really liked Illinois Capacitor (IC) because they were excellent and inexpensive. Looks like Cornell Dublier bought them. Cornell Dublier always made excellent capacitors, but they were pricey. I'm talking "back in the day" when we were selling new can multi-section capacitors.
Basically, I would buy with the OEM brand name on the capacitor. Capacitor winding machines and processes have come a long way, and they need maintenance. So I can't see anyone but a major capacitor manufacturer having that equipment, then putting someone else's name on it. That's the best scenario, the ugly would be a small company doing their best with old machines and little experience. If the major company is putting another name on the part, then the smaller brands are paying for that service and marking the capacitor up. I'm not sure about the economics, but I would think those might be more expensive than straight from the factory. Those will not be sold to industrial users or the aerospace / military complex. They wouldn't be able to satisfy the quality and delivery requirements anyway. That means their target market is audio or small low volume manufacturers (= high markup and low volume).
It's just like hearing something along the lines of "cheap 1N400x diodes", well they are highly engineered for the job they do, high reliability and made in the billions. That is why they are less expensive. Other rectifiers are made for a different purpose, have lower volumes and are therefore more expensive - not better. Of course you can buy "garbage" 1N400x diodes or anything else, but that has no bearing on the quality of a part made properly. Often using another rectifier actually doesn't work as well, it wasn't optimized for that job (60 Hz duty with reasonably sized filter capacitors). By the way, huge filter capacitors are not the way to go for a few reasons. Also, using a 1N400x rectifier for the wrong application would be just as silly (but I've seen it done). Then those don't perform well. Big surprise.
So choose parts designed for the job you're going to use them for. That takes some skill and knowledge, but don't blame the part if it was not used for what it was designed for. I think I just said, "there is no best capacitor".
-Chris
Basically, I would buy with the OEM brand name on the capacitor. Capacitor winding machines and processes have come a long way, and they need maintenance. So I can't see anyone but a major capacitor manufacturer having that equipment, then putting someone else's name on it. That's the best scenario, the ugly would be a small company doing their best with old machines and little experience. If the major company is putting another name on the part, then the smaller brands are paying for that service and marking the capacitor up. I'm not sure about the economics, but I would think those might be more expensive than straight from the factory. Those will not be sold to industrial users or the aerospace / military complex. They wouldn't be able to satisfy the quality and delivery requirements anyway. That means their target market is audio or small low volume manufacturers (= high markup and low volume).
It's just like hearing something along the lines of "cheap 1N400x diodes", well they are highly engineered for the job they do, high reliability and made in the billions. That is why they are less expensive. Other rectifiers are made for a different purpose, have lower volumes and are therefore more expensive - not better. Of course you can buy "garbage" 1N400x diodes or anything else, but that has no bearing on the quality of a part made properly. Often using another rectifier actually doesn't work as well, it wasn't optimized for that job (60 Hz duty with reasonably sized filter capacitors). By the way, huge filter capacitors are not the way to go for a few reasons. Also, using a 1N400x rectifier for the wrong application would be just as silly (but I've seen it done). Then those don't perform well. Big surprise.
So choose parts designed for the job you're going to use them for. That takes some skill and knowledge, but don't blame the part if it was not used for what it was designed for. I think I just said, "there is no best capacitor".
-Chris
With regards to diodes, any point in using fast, soft recovery in place of the 1N4003's?
I get where you are coming from with capacitors, I would have used solen 0.1uf simply because I keep a lot of their MP caps on hand, but I didnt have enough in the small size.
I get where you are coming from with capacitors, I would have used solen 0.1uf simply because I keep a lot of their MP caps on hand, but I didnt have enough in the small size.