Hi Donovas - the RC network is essentially an integrator, and contributes a dominant pole. The value of R isn't critical, and I've used values from zero (short) to about 1.5k in my prototypes - you should be fine with 470R to 1k. In simulation, it appears to flatten a kink in the closed-loop frequency response.
The value of C impacts the sonics and stability a bit more than R. The lower the value of C, the better the micro-detail, definition and air, but stability is compromised. The higher the value of C, the better the stability, but sonics become more nasal and metallic. In empirical practice, something between 22 pF and 47 pF appears to be the sweet spot, but will vary depending on the specific chipamp that's chosen.
It's also possible to add an interstage Zobel RC between the + and - inputs of the chipamp. I don't use one with either the LM1875 or the TDA2050, but if you feel the need to try a Zobel RC, a reasonable starting value is 3k3 + 1 nF in series, just like the MyRef Rev C. The interstage Zobel RC also contributes to stability, allowing the use of a smaller C for the global RC integrator above.
The value of C impacts the sonics and stability a bit more than R. The lower the value of C, the better the micro-detail, definition and air, but stability is compromised. The higher the value of C, the better the stability, but sonics become more nasal and metallic. In empirical practice, something between 22 pF and 47 pF appears to be the sweet spot, but will vary depending on the specific chipamp that's chosen.
It's also possible to add an interstage Zobel RC between the + and - inputs of the chipamp. I don't use one with either the LM1875 or the TDA2050, but if you feel the need to try a Zobel RC, a reasonable starting value is 3k3 + 1 nF in series, just like the MyRef Rev C. The interstage Zobel RC also contributes to stability, allowing the use of a smaller C for the global RC integrator above.
Hi
I’ve completed the kit earlier last month and tested in a stock 8 ohms 6” YP Speaker with the source from my Galaxy Note 2 mobile. Sound was really awesome; it’s very clean and natural even with this kind of cheap driver. There was no noise or humming present even without the proper grounding (star grounding I guess).
Though there was some problem initially but rectified it based on the Linuxguru’s suggestion. I really appreciate his support and guidance throughout the process and I would like to thanks Linuxguru for everything he has done for me.
Initially I’ve ordered a non-premium kit, but destroyed it along with major components. Then again I’ve ordered another premium kit and after listening to the accident he quoted me a lesser price without some components like TDA2050, Relay, Power Capacitors and some other small components which I’ve saved from the previous kit.
But I was really surprised after receiving the kit as there was everything including a pair of LM1875 (except the relay). Initially I thought that it might be mistakenly packed with all these components which he wasn’t supposed to send as I’ve not paid him for that. But when I was conveyed that, he said it was intentionally packed with all those components, even he denied to receive any further amounts for those extra components. He had also sent me a good potentiometer with the kit as I was unable to source it locally.
Here are some pictures of the completed kit…
I’ve completed the kit earlier last month and tested in a stock 8 ohms 6” YP Speaker with the source from my Galaxy Note 2 mobile. Sound was really awesome; it’s very clean and natural even with this kind of cheap driver. There was no noise or humming present even without the proper grounding (star grounding I guess).
Though there was some problem initially but rectified it based on the Linuxguru’s suggestion. I really appreciate his support and guidance throughout the process and I would like to thanks Linuxguru for everything he has done for me.
Initially I’ve ordered a non-premium kit, but destroyed it along with major components. Then again I’ve ordered another premium kit and after listening to the accident he quoted me a lesser price without some components like TDA2050, Relay, Power Capacitors and some other small components which I’ve saved from the previous kit.
But I was really surprised after receiving the kit as there was everything including a pair of LM1875 (except the relay). Initially I thought that it might be mistakenly packed with all these components which he wasn’t supposed to send as I’ve not paid him for that. But when I was conveyed that, he said it was intentionally packed with all those components, even he denied to receive any further amounts for those extra components. He had also sent me a good potentiometer with the kit as I was unable to source it locally.
Here are some pictures of the completed kit…
That is telling you you don't need star grounding.
The way you have it must be correct to give a Hum+Noise value that reads <0.1mVac.
Connect two interconnects and measure the H+N when the far ends are shorted and grounds/barrels cross coupled. It should still read VERY close to 0.0mVdc, maybe rising to <0.3mVac.
It should work with TO-3 and JFET input power opamps as well, but it may need tweaks to the compensation network for specific opamps. If stability is a problem, an additional interstage RC Zobel between the non-inverting and inverting inputs of the power opamp can help roll off some of the loop gain at HF. A good starting value is 3k3 in series with 1 nF, as in the MyRef Rev C. Generally, slower opamps in the outer loop are easier to stabilize - NE5532 and OPA2134 are excellent initial choices. If all else fails, TL072 should work.
I have no intention to change the outer loop opamp, just leave them be the 2134s.
I've read somewhere that a current-source amplifier could damage the speaker due to their effect on Qes parameter of the spker driver and they only work with specific type of speakers like full range...etc.Is it the case with MiniRef and MyRef ?
Thanks.
I've read somewhere that a current-source amplifier could damage the speaker due to their effect on Qes parameter of the spker driver and they only work with specific type of speakers like full range...etc.Is it the case with MiniRef and MyRef ?
Thanks.
Both the MyRef and MiniRef work fine with 2-way and 3-way speakers also, but they probably sound the best with fullrangers, with or without a small helper tweeter.
(I haven't heard of any report of speaker damage in normal operation with either amp, though it is entirely possible for a failed chipamp to short to either rail and damage the woofers in the brief instant before the speaker protection kicks in. However, that would typically happen only if the chipamp had inadequate heatsinking, was overloaded and/or overvolted, or was oscillating/unstable.)
(I haven't heard of any report of speaker damage in normal operation with either amp, though it is entirely possible for a failed chipamp to short to either rail and damage the woofers in the brief instant before the speaker protection kicks in. However, that would typically happen only if the chipamp had inadequate heatsinking, was overloaded and/or overvolted, or was oscillating/unstable.)
MyRef & MiniRef are only current-source internally. The speakers see them like any normal voltage source power amp.
__________________
.. with my speaker designer's hat on .. and off topic as MyRef & MiniRef are voltage sources.
It's not that current amps work only with full range speakers. Current drive modifies the frequency response of speakers (which are usually designed for voltage drive) and for full range speakers, this modified response usually sounds better .. more bass & less HF which most full range units need.
.. but you are just as likely to damage a speaker with a voltage source by turning up the bass control.
Hi Linuxguru,
Finally I made the Miniref,with the Board you gave me.
I had some trouble with fake IC which blew up with 18volt Supply and Sounded coarse and muddy with lesser Supply.Moreover TI was not supplying sample pieces for this IC anymore.Finally I purchased this chip through one of the Online DIY store from India.
When I first powered up there seemed to be strange Anomaly i.e output got muted frequently (not the relay) as if some kind of cut off or protection was triggered every 10 second or so.This muting happened between left and right channels variably.After much investigation I realized it could be only due to Overheating and Viola,the Chip Metal Tab was barely Touching the Heatsink.,Not unusual for To220 package but I missed it initially.
To make it seat properly on the heatsink I had to force bent the Chip backwards and used a clamp. And then, all well.
Right now powered with 20 volt AC Supply driving a 8ohm Loudspeaker.Playing directly from my Personal Computer without a volume control .
This cutie doesn't feel like an Under powered 20watts Amp. It have a lot of details and well managed low frequency response, Of course Not ground shaking but Enough to raise a discerning ear up to cloud nine.
Initial Listening made me feel bass is tight and deep with this amp compared to my Nmos200. I strongly feel the usual 'Presence' in the sound is absent with this Amp.
I have to do a lot of listening before arriving to a proper conclusion and I will be trying with different Loudspeakers and drivers.
I did not use the precision resistors instead managed with matched 1%Mfrs,It was tedious task matching the resistor ,it took me 3 years to finally reach somewhere with this thing
Assembling this board was cakewalk ,I would say A perfect Amplifier Kit with Thermal protection and Output turn on delay.
Regards.
Finally I made the Miniref,with the Board you gave me.
I had some trouble with fake IC which blew up with 18volt Supply and Sounded coarse and muddy with lesser Supply.Moreover TI was not supplying sample pieces for this IC anymore.Finally I purchased this chip through one of the Online DIY store from India.
When I first powered up there seemed to be strange Anomaly i.e output got muted frequently (not the relay) as if some kind of cut off or protection was triggered every 10 second or so.This muting happened between left and right channels variably.After much investigation I realized it could be only due to Overheating and Viola,the Chip Metal Tab was barely Touching the Heatsink.,Not unusual for To220 package but I missed it initially.
To make it seat properly on the heatsink I had to force bent the Chip backwards and used a clamp. And then, all well.
Right now powered with 20 volt AC Supply driving a 8ohm Loudspeaker.Playing directly from my Personal Computer without a volume control .
This cutie doesn't feel like an Under powered 20watts Amp. It have a lot of details and well managed low frequency response, Of course Not ground shaking but Enough to raise a discerning ear up to cloud nine.
Initial Listening made me feel bass is tight and deep with this amp compared to my Nmos200. I strongly feel the usual 'Presence' in the sound is absent with this Amp.
I have to do a lot of listening before arriving to a proper conclusion and I will be trying with different Loudspeakers and drivers.
I did not use the precision resistors instead managed with matched 1%Mfrs,It was tedious task matching the resistor ,it took me 3 years to finally reach somewhere with this thing
Assembling this board was cakewalk ,I would say A perfect Amplifier Kit with Thermal protection and Output turn on delay.
Regards.
Last edited:
Congrats!!!!
You can try rolling different opamp for different signature.
I've tried with opa2134, tl072, ne5532 and of course lt1361 and found it sounds best to my ear with LT1361. It's much transparent and crisp.
But to get best out of it you should try linuxguru's lf02 module with slow opamp.
Tough In my case the lf02 is having a relatively fast opamp so it became a bit unstable but beleave me the quality of sonic was just out of the world with the module.
A small review on LF02 module:
It's a great sounding module. I've never found anything like this before. The sonics are worth dying for. Its like I'm using an equalizer with some automatic best setting for each and every track. Crisp and crystal like sound. Can easily identify each and every instruments. Even I found the presence of some instruments which I've never heard before on those track. In words I can't describe the experience. Moreover the lower frequencies are much better defined with it.
Sent from my GT-N7100 using Tapatalk
Last edited:
The Amp is proving to be worth every effort ,it is unbelievably cheap for the performance.
What is this magic,why it sounds different than rest of the standard AB class amp.Is it just a Psychological effect?
What if we have used the Output Amp as mere unity buffer instead of Howland current pump design.
I could not figure out what is happening in this.
Also note: I don't see any need of Ultra fast diodes for AC rectification,I did not feel any different no matter how hard I tried by replacing it with a 4 pin standard rectifier.But I reverted to UF diodes as of now.
Changing the 5532 to OPA2134 made some difference to my listening or was it just a placebo effect.
NE5532 had a very little bit of hiss feel though unheard when Idle but I felt it in deep silence of mid night while silent gaps of music or the Harmonics or was I too sleepy ?
It is very difficult to compare 2 amps simultaneously Probably A switching box with DPDT switch could prove helpful, does it?
How is such detailed listening carried out in normal living rooms.
Regards.
What is this magic,why it sounds different than rest of the standard AB class amp.Is it just a Psychological effect?
What if we have used the Output Amp as mere unity buffer instead of Howland current pump design.
I could not figure out what is happening in this.
Also note: I don't see any need of Ultra fast diodes for AC rectification,I did not feel any different no matter how hard I tried by replacing it with a 4 pin standard rectifier.But I reverted to UF diodes as of now.
Changing the 5532 to OPA2134 made some difference to my listening or was it just a placebo effect.
NE5532 had a very little bit of hiss feel though unheard when Idle but I felt it in deep silence of mid night while silent gaps of music or the Harmonics or was I too sleepy ?
It is very difficult to compare 2 amps simultaneously Probably A switching box with DPDT switch could prove helpful, does it?
How is such detailed listening carried out in normal living rooms.
Regards.
Infact I am utterly confused, some one please teach me how to do a detailed listening before I proceed with Lf02 module.
Theoretically I believe it is impossible to detect the sonic flaws from Opamp like NE5532,it is a respectably a good opamp.
Then the mild difference,is it due to design,where faster opamps are needed,or the some unnoticed harmonics or oscillations?
Verdict: Miniref sounds great !
I am planning to use Full range speakers for this Amp.Before that I wish to add some active frequency correction filter to optimize the full range bumps and dips.
Would this filter degrade the charecteristics of Miniref amp?
I planned to use Ne5532 in the filter with unity gain .
If you can reliably pick up differences in a DBLT, I'll put a LOT of money that it's oscillations.
In my previous life, I was a DBLT guru and ALL the differences picked out by my Listening Panel (which had some of the best ears in the business) were due to oscillations.