*Did you intend “10,000” instead of “1000?” Otherwise “15000 is a BIG jump! Increasing from 10K to 12K or 15K is not unusual w/out taxing the PS components.
**What does “4-pin shape” mean?
Thanks. Tony
Per,
Advice for non-negotiable test equipment to have on hand to commence tweaking RB-981 power amps?
I have two DMMs - one more accurate than the other. I’m confident you’ll tell me “add an “O” scope” (suggestions please?) to do even some the simpler stuff like PS cap* upgrades* and tune/tweak bias settings after replacing the pots.
*like Paulus
But I won’t stop there as the 981 & 991 sport nearly identical front ends.
I have 2 spares. I plan to do one at a time w/focus on what makes for a sweeter mid & top end.
A 1000 thanks in advance. Tony
Advice for non-negotiable test equipment to have on hand to commence tweaking RB-981 power amps?
I have two DMMs - one more accurate than the other. I’m confident you’ll tell me “add an “O” scope” (suggestions please?) to do even some the simpler stuff like PS cap* upgrades* and tune/tweak bias settings after replacing the pots.
*like Paulus
But I won’t stop there as the 981 & 991 sport nearly identical front ends.
I have 2 spares. I plan to do one at a time w/focus on what makes for a sweeter mid & top end.
A 1000 thanks in advance. Tony
Sorry, a lot of typos were made in the previous post.
Yes, I meant the upgrade from 10.000uF to 15k or 22k, since the only these can be find in my area.
In RA-1070 the power supply capacitors are not standard 2-pin snap in, but 4-pin type, thus making a replacement with 2-pin capacitors somewhat more complicated. And the only caps I can pouchase are those listed above. Kemet (which own now BHC) lists 4-pin caps, but since these are actually same as are in Rotel amps (Rotel-rebranded BHC), I wish to change to some other brand (Nichicon, United Chemi-Con, BC components...) which are more transparent and smooth in my previous experience. But...the only brand I can find here where I'm living are CDE 22000uF 4-pin.
Gee, Tony...
I honestly hesitate to give a straight answer to your questions.
It depends on whether or not you 'get the bug' to start experimenting with amps and circuits. Then of course you would start needing an audio generator, scope etc. There is now quite a range of reasonably affordable stuff available online these days.
For the work I plan to do on the RA-971 I'll definitely start with using external +/- 30V power supplies with current limiters to avoid serious burnouts while fiddling with things.
But come to think of it, when I do the current standard CM1 and VAS3 upgrades for customers I rarely use anything but my trusted old Fluke 87 DMM (because everything - if done correctly - normally works out just fine) - then followed by a RTA measurement under REW, USB connected to a separate 8ohm power load box with a XonarU7 soundcard.
I did and will put a lot of R&D effort into ensuring that the CM and VAS modules work safely and with good thermal and HF stability in their applications so that upgraders (including myself) relatively easily can get things done without any heavy investments.
Now, for a 'sweeter' mid & top end - I'm not quite sure what you mean. The upgrade will give you a much more detailed, punchy and precise soundstage while taking away the 'soft/pleasant' Rotel sound.
Simply by a 20dB reduction of the shoal of distortion piranhas that munch away at good music reproduction.
Please note that I don't dislike distortion per se, but I much prefer to listen directly to that from, say, Santana's guitar amp's than any additions from my own amp.
Hope this rather bumbling reply helps.
Per
Last edited:
Solved: Meridian 557 Power-on Problem
It turned out that the problem was age-failure of the caps on the low-current +15v & -15v power circuits and the bias circuit including C43, C3, C18. I also replaced C4, C6, C20 and C24 since they were starting to show age failure trend along with a few other of the 10uf caps in the same area just to be safe. C43 was the primary culprit causing the shut down. I used Nichicon Fine Gold caps as Meridian did in the original design. There was a very noticeable improvement in the sound.
The big high-current power supply caps were exactly to spec and did not need to be replaced.
I still have a question regarding whether the 557 amp could benefit from adding C11 and C13 into the circuit. There is a place for these caps on the left and right channel circuit boards but they were left out probably because they wouldn't fit due to the large toroidal transformer being in the way. The two big caps that are installed in the PCB (C32 & C54) are at the V+ and V- buses at the power MOSFETS while the missing C11 and C13 are on V+ and V- at the outputs of the bridge rectifiers. There are caps available per spec that will fit now since they can make them smaller for a given capacity. Does anyone think it's worth installing them since they would be essentially be in parallel? However, from a dynamic power cycle perspective I assume that C11 and C13 would get the primary charge load from the power supply while C32 and C54 would get the primary discharge load based on the output power being pulled by the MOSFETS if all were installed.
D
It turned out that the problem was age-failure of the caps on the low-current +15v & -15v power circuits and the bias circuit including C43, C3, C18. I also replaced C4, C6, C20 and C24 since they were starting to show age failure trend along with a few other of the 10uf caps in the same area just to be safe. C43 was the primary culprit causing the shut down. I used Nichicon Fine Gold caps as Meridian did in the original design. There was a very noticeable improvement in the sound.
The big high-current power supply caps were exactly to spec and did not need to be replaced.
I still have a question regarding whether the 557 amp could benefit from adding C11 and C13 into the circuit. There is a place for these caps on the left and right channel circuit boards but they were left out probably because they wouldn't fit due to the large toroidal transformer being in the way. The two big caps that are installed in the PCB (C32 & C54) are at the V+ and V- buses at the power MOSFETS while the missing C11 and C13 are on V+ and V- at the outputs of the bridge rectifiers. There are caps available per spec that will fit now since they can make them smaller for a given capacity. Does anyone think it's worth installing them since they would be essentially be in parallel? However, from a dynamic power cycle perspective I assume that C11 and C13 would get the primary charge load from the power supply while C32 and C54 would get the primary discharge load based on the output power being pulled by the MOSFETS if all were installed.
D
Hi,
Yes, it is a very good amp, using the same power stage as the 991, etc.You can find both user and service manuals on https://www.vintageshifi.com/repertoire-pdf/Rotel.php
There seems to be quite some interest on possible upgrades on these amps, and I will get around to do some work and post details on this as and when.
Best,
Per
Hi S Sam,
This 330pF // 330k brutal lag to ground is used by Curtis in all Rotel amps regardless of the topology.
It reduces OL HF gain to ensure stability, but also introduces the distortion known as the 'warm' Rotel sound.
You could of course divide it up (165pF // 660k - or nearest) to each side of the bias circuit - mainly to please the schematic appearance.
But a 100kHz signal will simply see these two loads as in parallel, just separated by a ~300 Ohm bias circuit resistance - further bypassed by 10uF! And not give a hoot.
Best,
Per
This 330pF // 330k brutal lag to ground is used by Curtis in all Rotel amps regardless of the topology.
It reduces OL HF gain to ensure stability, but also introduces the distortion known as the 'warm' Rotel sound.
You could of course divide it up (165pF // 660k - or nearest) to each side of the bias circuit - mainly to please the schematic appearance.
But a 100kHz signal will simply see these two loads as in parallel, just separated by a ~300 Ohm bias circuit resistance - further bypassed by 10uF! And not give a hoot.
Best,
Per
Hi there,
Try DNM Design for caps, stands for Dennis Neal Morecroft, he invented the Slit Foil cap used in Rotel amps.
I bought them to and wasn't disappointed
He also sells the 4 pin caps
www dot dnm dot co dot uk
Cheers, Paulus
Last edited:
@Smartdriver
Thank you for reply. I am aware of DNM capacitors, but the maximum capacitans listed are 10000uF, so I can not increase the total capacitanse as intended.
And, to be honest, I cannot be sure these in Rotel are DNM "T network" capacitors, or simply 4-pin standard capacitors. On the board, I found (also in schematiscs) printed sgnature under the caps as "in" and "out", but again, cannot be quite sure about this.
If so, I really don't know if there can be replaced DNM with ordinary 4-pin caps.
Maybe some of the fellows here can clarified this?
Thank you for reply. I am aware of DNM capacitors, but the maximum capacitans listed are 10000uF, so I can not increase the total capacitanse as intended.
And, to be honest, I cannot be sure these in Rotel are DNM "T network" capacitors, or simply 4-pin standard capacitors. On the board, I found (also in schematiscs) printed sgnature under the caps as "in" and "out", but again, cannot be quite sure about this.
If so, I really don't know if there can be replaced DNM with ordinary 4-pin caps.
Maybe some of the fellows here can clarified this?
@eradoncic;
You could try and contact HANEX service by phone or Email
Rotel - Hanex Service | Audio Repairs, Service and Maintenance
They are able to help you with your question about the 4 pin or T network caps in your Rotel amp.
I had the same problem when changing psu caps, stock is 15K and only 10K aftermarket so i doubled up.
I paralleled 2 x 10K to get 20K
Cheers, Paulus
You could try and contact HANEX service by phone or Email
Rotel - Hanex Service | Audio Repairs, Service and Maintenance
They are able to help you with your question about the 4 pin or T network caps in your Rotel amp.
I had the same problem when changing psu caps, stock is 15K and only 10K aftermarket so i doubled up.
I paralleled 2 x 10K to get 20K
Cheers, Paulus
To Per, Paulus, and anyone w/PS upgrade knowledge/experience:
the smallest DNM Design sells are 10KuF - the same size as my Rotels. I think doubling filtering capacitance to 20KuF may be too much for the transformer but I’m far from an expert in these matters.
Side note to Per: in your test equipment reply you mentioned an RTA & REW. I have a Behringer DEQ2486 w/an RTA and it’s mic for the the same. Would that suffice. I’ve used it to tame peaks. In the process, I discovered & corrected flipped polarity to my tweeters.
Many thanks for you help & patience. Tony.
suggestions for alternative 63V PS caps smaller than 10KuF?Hi there,
Try DNM Design for caps, stands for Dennis Neal Morecroft, he invented the Slit Foil cap used in Rotel amps.
I bought them to and wasn't disappointed
He also sells the 4 pin caps
www dot dnm dot co dot uk
Cheers, Paulus
the smallest DNM Design sells are 10KuF - the same size as my Rotels. I think doubling filtering capacitance to 20KuF may be too much for the transformer but I’m far from an expert in these matters.
Side note to Per: in your test equipment reply you mentioned an RTA & REW. I have a Behringer DEQ2486 w/an RTA and it’s mic for the the same. Would that suffice. I’ve used it to tame peaks. In the process, I discovered & corrected flipped polarity to my tweeters.
Many thanks for you help & patience. Tony.
Hi Tony,
It is true that doubling of the caps to 20,000uF will also double the 'inrush' current duration and thus thermal load on the trafo and Graetz diode bridge, but that is something these components should be able to withstand. The trafo may just go into saturation a second or two longer and then back out into normal operation. I wouldn't worry.
Not familiar with the DEQ2486, but I guess that like the REW it is not up to, say, Audio Precision standard (or price!). But all you need is for it to to show whether there is any improvement or not from the upgrades.
Best,
Per
It is true that doubling of the caps to 20,000uF will also double the 'inrush' current duration and thus thermal load on the trafo and Graetz diode bridge, but that is something these components should be able to withstand. The trafo may just go into saturation a second or two longer and then back out into normal operation. I wouldn't worry.
Not familiar with the DEQ2486, but I guess that like the REW it is not up to, say, Audio Precision standard (or price!). But all you need is for it to to show whether there is any improvement or not from the upgrades.
Best,
Per
Hi there Tony,
Why would you go lower then the stock 10KuF
I'm with Per on this but i've also read somewhere on this forum that there is a "limit" in increasing value of the psu caps.
The drawback is that for mentioned inrush current!
If the psu caps don't get enough time to charge because they are so big then they can not do there job properly.
If your psu caps aren't fully charged they do not deliver to specification.
This could be audible but I don't know for sure
I know, I did this, but in defence, I went from 15K to 20K, a third, you are doubling up.
I guess that it depends on the capacity of the transformer and rectifier.
If they are able to produce a large amount of power and handle enough current then there should be no problem.
Don't know if I made anything clear here but I tried
@smartdriver
I agree with you, good power supply consists either excellent transformer (and less capacitance) or average transformer and more capacitors.
For example: some Krell and Sansui best amps with big (and probably good good regulated transformer) usually have unexpected low capacitance (10000uF or even less), and on the other hand - big bank is more frequenty find in less quality transformer (standard toroidal for example)
I feel that rotels are in the middle of this extremes, but being also transformer manufactorers for own products - there is above average PS pushed a bit to upper limits.
I was told all this by enginers, I'm not qualified in electroncs, so don't take this as a fact.
I agree with you, good power supply consists either excellent transformer (and less capacitance) or average transformer and more capacitors.
For example: some Krell and Sansui best amps with big (and probably good good regulated transformer) usually have unexpected low capacitance (10000uF or even less), and on the other hand - big bank is more frequenty find in less quality transformer (standard toroidal for example)
I feel that rotels are in the middle of this extremes, but being also transformer manufactorers for own products - there is above average PS pushed a bit to upper limits.
I was told all this by enginers, I'm not qualified in electroncs, so don't take this as a fact.
The bulk capacitance is required to provide power during the gaps between the mains waveform peaks and is little to do with the quality of the transformer - its all about the load requirement for current during this period - this means that all you need to know is the mains frequency and impedance of the load and you can determine how much capacitance needed (for a given percentage droop across a half-wave).
The quality of the transformer affects the voltage droop at high load too, so there is some tradeoff you can make between bulk capacitance and transformer efficiency, but its pretty minor - you can't really make up for real inadequacies in one by increasing the other, both should be adequate to start with.
In commercial equipment the transformer is usually undersized for continuous full power as this is seldom seen in music signals. However the caps need to be able to hold up across a half-cycle during short bursts of power, and shouldn't be undersized.
As an example of a calculation for a capacitor, consider an 8 ohm load, 50Hz mains, and a desire for only 5% droop across a half-cycle. This means the RC time-constant should be 0.01s / -ln(0.95) = 195ms, so C = 0.195/8 = 24mF.
However the half-cycle delay is somewhat less than half the period, and 5% droop is only 0.4dB or so, and the actual impedance seen by the supply is more than the loudspeaker load due to voltage across the output transistors, so this is rather over-engineered in practice.
The quality of the transformer affects the voltage droop at high load too, so there is some tradeoff you can make between bulk capacitance and transformer efficiency, but its pretty minor - you can't really make up for real inadequacies in one by increasing the other, both should be adequate to start with.
In commercial equipment the transformer is usually undersized for continuous full power as this is seldom seen in music signals. However the caps need to be able to hold up across a half-cycle during short bursts of power, and shouldn't be undersized.
As an example of a calculation for a capacitor, consider an 8 ohm load, 50Hz mains, and a desire for only 5% droop across a half-cycle. This means the RC time-constant should be 0.01s / -ln(0.95) = 195ms, so C = 0.195/8 = 24mF.
However the half-cycle delay is somewhat less than half the period, and 5% droop is only 0.4dB or so, and the actual impedance seen by the supply is more than the loudspeaker load due to voltage across the output transistors, so this is rather over-engineered in practice.
Thankx for you reply Mark, but ...
Krell KAV-300I is rated as 2x150/300W 8/4 Ohms
It's transformer is rated 400VA (as per reviews), and there is only 2x 8200uF/80V per channel; his "descent" KAV 400ix, has transformer rated 750VA and 4x6800uF/63V capacitor bank per channel. I know this since I had to replace these caps and there is no chance to increase even for 1000uF because of unusually limited space, just underneath preamp which literally sit on the bank - and all are 35x30mm (w/h)!! The amp started do have "short breath" on bass, and all caps measured less than 5000uF.
So, something here is out of ordinary, at least, so to say.
Further more, Sansui 907MRX and several other Alphas approx 600-750VA rating transformer and just 5600+3900uF /85V capacitors per channel (yes, they are assymetrical). For 160W/ch !
Confusing, anyway.
Krell KAV-300I is rated as 2x150/300W 8/4 Ohms
It's transformer is rated 400VA (as per reviews), and there is only 2x 8200uF/80V per channel; his "descent" KAV 400ix, has transformer rated 750VA and 4x6800uF/63V capacitor bank per channel. I know this since I had to replace these caps and there is no chance to increase even for 1000uF because of unusually limited space, just underneath preamp which literally sit on the bank - and all are 35x30mm (w/h)!! The amp started do have "short breath" on bass, and all caps measured less than 5000uF.
So, something here is out of ordinary, at least, so to say.
Further more, Sansui 907MRX and several other Alphas approx 600-750VA rating transformer and just 5600+3900uF /85V capacitors per channel (yes, they are assymetrical). For 160W/ch !
Confusing, anyway.