LM1875 nested Gainclone

[linuxguru]

My apologies! NE5532 has not yet arrived.

The reputable JRC NJM4560 did arrive and I could employ them if you think they are compatible with this project.

There's probably no harm in trying out the 4560 instead of the the NE5532 here. Is it also a dual op-amp in a standard 8-pin package?

It's probably a bit faster than the NE5532, but the compensation scheme should take care of that - though the sonics may not be optimal. You may want to go with lower +/- Vs rails initally, and check for HF oscillations with an oscilloscope before going to 28 or 35 V rails.

 

[danielwritesbac]

There's probably no harm in trying out the 4560 instead of the the NE5532 here. Is it also a dual op-amp in a standard 8-pin package?

It's probably a bit faster than the NE5532, but the compensation scheme should take care of that - though the sonics may not be optimal. You may want to go with lower +/- Vs rails initally, and check for HF oscillations with an oscilloscope before going to 28 or 35 V rails.

NXP NE5532, 10mhz, Slew9V, unity stable
JRC NJM4560, 10mhz, Slew4V
*Pin compatible
**4560L isn't pin compatible, but rather 8 pins straight in a row, and its rated to 800mw PD although it can do 1 watt, in this larger version--I have these.

 

[aparatusonitus]

Thanks to Pedja for his LM1875 model.

Hi linuxguru,

Could you please upload the LM1875 LTspice model...I have the Modpex model, but it's spice3 format and I'm not smart enough to convert it to psice format, thank you.

 

[linuxguru]

Could you please upload the LM1875 LTspice model...I have the Modpex model, but it's spice3 format and I'm not smart enough to convert it to psice format, thank you.

Here it is:

http://www.diyaudio.com/forums/chip-amps/138762-lm1875-nested-gainclone-3.html#post1813245

Pedja probably converted it to PSpice. No guarantees about its accuracy - it seems to have far more diodes and than transistors. It seems to pass sanity checks when used in simple inverting and non-inverting configurations.

 

[aparatusonitus]

Hi linuxguru,

Thank you for replay...It occurred to me that model is the same LM1875 spice model that I've modded to use with LTspice after Pedja showed me how to do that...it's seems I'm loosing my brain cells rapidly 8-(

For reference and if one want to tweak it...

.SUBCKT LM1875 3 2 6 4 5
* EXTERNAL NODE DESIGNATIONS
* 5 --> - SUPPLY (VEE)
* 4 --> + SUPPLY (VCC)
* 6 --> OUTPUT
* 2 --> (-) INPUT
* 3 --> (+) INPUT
EOX 120 10 31 32 2.0
RCX 120 121 1K
RDX 121 10 1K
RBX 120 122 1K
GOS 10 57 122 121 1.0
RVOS 31 32 1K
RINB 2 18 1000
RINA 3 19 1000
DIN1 5 18 DMOD2
DIN2 18 4 DMOD2
DIN3 5 19 DMOD2
DIN4 19 4 DMOD2
EXX 10 5 17 5 1.0
EEE 10 50 17 5 1.0
ECC 40 10 4 17 1.0
RAA 4 17 100MEG
RBB 17 5 100MEG
ISET 10 24 1e-3
DA1 24 23 DMOD1
RBAL 23 22 1000
ESUPP 22 21 4 5 1.0
VOFF 21 10 -1.25
DA2 24 25 DMOD1
VSENS1 25 26 DC 0
RSET 26 10 1K
CSET 26 10 1e-10
FSET 10 31 VSENS1 1.0
R001 34 10 1K
FTEMP 10 27 VSENS1 1.0
DTA 27 10 DMOD2
DTB 28 29 DMOD2
VTEMP 29 10 DC 0
ECMR 38 10 11 10 1.0
VCMX 38 39 DC 0
RCM2 41 10 1MEG
EPSR 42 10 4 10 1.0
CDC1 43 42 10U
VPSX 43 44 DC 0
RPSR2 45 10 1MEG
FCXX 57 10 VCXX 100
DCX1 98 97 DMOD1
DCX2 95 94 DMOD1
RCX1 99 98 100
RCX2 94 99 100
VCXX 99 96 DC 0
ECMX 96 10 11 10 1.0
DLIM1 52 57 DMOD1
DLIM2 57 51 DMOD1
ELIMP 51 10 26 10 99.3
GDM 10 57 3 2 1
C1 58 59 1e-10
DCLMP2 59 40 DMOD1
DCLMP1 50 59 DMOD1
RO2 59 10 1K
GO3 10 71 59 10 1
RO3 71 10 1
DDN1 73 74 DMOD1
DDN2 73 710 DMOD1
DDP1 75 72 DMOD1
DDP2 71 720 DMOD1
RDN2 710 71 100
RDP 720 72 100
VOOP 40 76 DC 0
VOON 77 50 DC 0
QNO 76 73 78 NPN1
QNP 77 72 79 PNP1
RNO 78 81 1
RPO 79 81 1
VOX 86 6 DC 0
RNT 76 81 100MEG
RPT 81 77 1MEG
FX 10 93 VOX 1.0
DFX1 93 91 DMOD1
VFX1 91 10 DC 0
DFX2 92 93 DMOD1
VFX2 10 92 DC 0
FPX 4 10 VFX1 1.0
FNX 10 5 VFX2 1.0
RAX 122 10 MRAX 1.002000e+03
.MODEL MRAX RES TC1=5e-05
FIN1 18 5 VTEMP 0.875
FIN2 19 5 VTEMP 1.125
CIN1 2 10 1e-12
CIN2 3 10 1e-12
RD1 18 11 500000
RD2 19 11 500000
RCM 11 10 9.75e+06
FCMR 10 57 VCMX 31.6228
FPSR 10 57 VPSX 63.2456
RSLOPE 4 5 1000
GPWR 4 5 26 10 0.02
ETEMP 27 28 32 33 0.219999
RIB 32 33 MRIB 1K
.MODEL MRIB RES TC1=0.00293599
RISC 33 34 MRISC 1K
.MODEL MRISC RES TC1=-0.003
RCM1 39 41 316.228
CCM 41 10 1.59155e-10
RPSR1 44 45 10000
CPSR 45 10 5.30516e-10
ELIMN 10 52 26 10 99.3
RDM 57 10 420.038
C2 57 10 3.4446e-11
ECMP 40 97 26 10 0.5
ECMN 95 50 26 10 0.5
G2 58 10 57 10 9.5e-06
R2 58 10 250.604
GO2 59 10 58 10 32
EPOS 40 74 26 10 0
ENEG 75 50 26 10 0.1
GSOURCE 74 73 33 34 0.04
GSINK 72 75 33 34 0.04
ROO 81 86 0.5
.MODEL DMOD1 D
*-- DMOD1 DEFAULT PARAMETERS
*IS=1e-14 RS=0 N=1 TT=0 CJO=0
*VJ=1 M=0.5 EG=1.11 XTI=3 FC=0.5
*KF=0 AF=1 BV=inf IBV=1e-3 TNOM=27
.MODEL DMOD2 D (IS=1e-17)
*-- DMOD2 DEFAULT PARAMETERS
*RS=0 N=1 TT=0 CJO=0
*VJ=1 M=0.5 EG=1.11 XTI=3 FC=0.5
*KF=0 AF=1 BV=inf IBV=1e-3 TNOM=27
.MODEL NPN1 NPN (BF=100 IS=1e-15)
*-- NPN1 DEFAULT PARAMETERS
*NF=1 VAF=inf IKF=inf ISE=0 NE=1.5
*BR=1 NR=1 VAR=inf IKR=inf ISC=0
*NC=2 RB=0 IRB=inf RBM=0 RE=0 RC=0
*CJE=0 VJE=0.75 MJE=0.33 TF=0 XTF=0
*VTF=inf ITF=0 PTF=0 CJC=0 VJC=0.75
*MJC=0.33 XCJC=1 TR=0 CJS=0 VJS=0.75
*MJS=0 XTB=0 EG=1.11 XTI=3 KF=0 AF=1
*FC=0.5 TNOM=27
.MODEL PNP1 PNP (BF=100 IS=1e-15)
*-- PNP1 DEFAULT PARAMETERS
*NF=1 VAF=inf IKF=inf ISE=0 NE=1.5
*BR=1 NR=1 VAR=inf IKR=inf ISC=0
*NC=2 RB=0 IRB=inf RBM=0 RE=0 RC=0
*CJE=0 VJE=0.75 MJE=0.33 TF=0 XTF=0
*VTF=inf ITF=0 PTF=0 CJC=0 VJC=0.75
*MJC=0.33 XCJC=1 TR=0 CJS=0 VJS=0.75
*MJS=0 XTB=0 EG=1.11 XTI=3 KF=0 AF=1
*FC=0.5 TNOM=27
.ENDS lm1875

 

[danielwritesbac]

Question:
Are authentic Philips NE5532AN painted or etched (mine are painted moniker)? And, is $5 with free shipping an "impossible" price for 10 units?

I'm thinking that it sounded like a good deal to begin with, but maybe there is an unexpected "OEM" op-amp inside these, and so I might have purchased the wrong part. Any clues?

 

[linuxguru]

Question:
Are authentic Philips NE5532AN painted or etched (mine are painted moniker)? And, is $5 with free shipping an "impossible" price for 10 units?

Here I can only get TI parts, and they seem to be lightly etched. $0.50 each is perfectly possible - I pay about $0.12 for my TI parts (which are no different sonically, IMHO), so there's no harm in paying say $0.20 each for NOS Signetics/Philips parts + $3 shipping = $5. Entirely reasonable...

However, if the source is an EBay seller in HK or China, a bit of vigilance is in order.

 

[danielwritesbac]

Here I can only get TI parts, and they seem to be lightly etched. $0.50 each is perfectly possible - I pay about $0.12 for my TI parts (which are no different sonically, IMHO), so there's no harm in paying say $0.20 each for NOS Signetics/Philips parts + $3 shipping = $5. Entirely reasonable...

However, if the source is an EBay seller in HK or China, a bit of vigilance is in order.

I'm sorry to report that Signetics nos advertisement is a sure fire way to get a very underwhelming fake op-amp at an inflated price. If operated as a buffer, it will play only its inbuilt song: "flip-a-flip-a-flip-a. . ."

Right now, the only authentic and very stable op-amp I have available is TL061. 4580's haven't arrived yet. I could try again to get some real NE5532's--any suggestions?

Question:
With NE5532 and 4580, stereo op-amps feeding single power op-amps, why not use the NE5532's or njm4580's 2 channels in parallel to turn some noise into a tiny bit of heat? Just a curiosity, but I noticed that Each channel on the Revo 5.1's silver 1/8" jack has a stereo op-amp, and there are four good size "47R5" resistors right before the jack. That sounds great! But, the green jack with just one stereo op-amp is lackluster and just downright boring. So, what do you think of parallel?

 

[linuxguru]

.
Question:
With NE5532 and 4580, stereo op-amps feeding single power op-amps, why not use the NE5532's or njm4580's 2 channels in parallel to turn some noise into a tiny bit of heat?

It will increase the current drive of the op-amp, for sure, but in this topology the NE5532 is not pushing too much current (a few mA at most) at any rate. I don't think there's any benefit here.

OTOH, for a headphone amp, paralleling op-amps outputs is a generally a sound idea.

 

[danielwritesbac]

It will increase the current drive of the op-amp, for sure, but in this topology the NE5532 is not pushing too much current (a few mA at most) at any rate. I don't think there's any benefit here.

OTOH, for a headphone amp, paralleling op-amps outputs is a generally a sound idea.

Well, I'm curious. If it isn't pushing much current, then is there an advantage to getting it to push some current, as in partial biasing to class-A?

 

[linuxguru]

Partial Class-A biasing of the input op-amp might be worth considering - I looked at it a while ago in another context, and it did not make much difference in the sim then. However, for this topology it might help the sonics a bit. I need a structural (i.e. transistor-level) model of the input op-amp to model this right.

 

[sfthurber]

Question:

... which makes the power circuit for LM3886 look like this so that LM3886 can run cool enough to drive a 4 ohm speaker from a single chip:

AC
Transformer
Rectifier
2200uF per rail
0.1R per rail (R16, R17)
--optional power supply interconnect--
10,000uF per rail (or 5x 2200uF per rail)
100uF polyester cap per rail
3.3uF (or 4.7uF) high esr polyester cap V+ to V-
Star Ground, Speaker Ground, 0V
Cable (not long)
amp board
220uF//220uF per rail
100nF ceramic or polypropylene cap per rail
lm3886
4 ohm speaker (or 16 ohm speaker or 8 ohm speaker)

This works in practice to get more than enough thermal margin for TDA7294 and LM3886 to run 4 ohm speaker per each chip. Cleaner power makes for cooler heatsink--no extraneous heating like the 47 labs style amps.
....

this is really interesting. do you mind if i ask you to clarify one point? Does "Star Ground, Speaker Ground, 0V" mean that the three are tied at the amp end of the cable or run separately in the cable and tied at the PSU end?

 

[AudioLapDance]

OPA604 & TDA7293

How about OPA604 & TDA7293?



OPA604:
LOW DISTORTION: 0.0003% at 1kHz
LOW NOISE: 10nV/√Hz
HIGH SLEW RATE: 25V/μs
WIDE GAIN-BANDWIDTH: 20MHz
UNITY-GAIN STABLE
WIDE SUPPLY RANGE: VS = ±4.5 to ±24V
DRIVES 600Ω LOAD
DUAL VERSION AVAILABLE (OPA2604)
MAX +/- 25V
http://www.ti.com.cn/general/cn/docs/lit/getliterature.tsp?genericPartNumber=opa604&fileType=pdf





TDA7293

VERY HIGH OPERATING VOLTAGE RANGE (±50V)
DMOS POWER STAGE
HIGH OUTPUT POWER (100W @ THD = 10%, RL = 8W, VS = ±40V)
MUTING/STAND-BY FUNCTIONS
NO SWITCH ON/OFF NOISE
VERYLOW DISTORTION
VERYLOW NOISE
SHORT CIRCUIT PROTECTION
THERMAL SHUTDOWN
CLIP DETECTOR
MODULARITY (MORE DEVICES CAN BE EASILY CONNECTED IN PARALLEL TO DRIVE VERY LOW IMPEDANCES)
View attachment TDA7293-1.pdf




Hmmm, 604 offset trim pins are ~ before / after "distortion rejection circuitry" ...

Hmmm mmmm, 7293 has separate input section and is easily 'stack-able' (better way: master and multiple slaves) for very high current to go along with its high voltage (±40V, test MAX ±50V, abs MAX ±60V) and has the buffer driver pin and boot loader / bootstrap pins

 

[AndrewT]

Follow Jung's rules for a composite opamp.
It should make for a very good signal and that may mean very good sound.

BTW
quoting 100W for +-40Vdc supply rails shows you are not interested in a good signal and certainly not interested in obtaining a good sound.

 

[papasteack]

Hi,
I lack of knowledge, and it's so maybe off topic, but isn't this design http://www.diyaudio.com/forums/chip-amps/233411-batomm-amp-lm-bridge-consider-try.html a neasted lm3886/lme49720 bridged remote sensing amp ? It seems to work very well from what i've google traducted from it serbian forum...
Thanks for clarifying if i'm wrong

Damien
http://www.diyaudio.com/forums/chip-amps/233411-batomm-amp-lm-bridge-consider-try.html

 

[carlmart]

Follow Jung's rules for a composite opamp.
It should make for a very good signal and that may mean very good sound.

Interesting. How would you go about it? Same values as on the original combo?

What about DC offset? It probably needs some feedback decoupling on one or both chips.

It would be interesting to try it with an LM3886.

And it would be great if Walt was reading this and contributed some tip to go about it.

 

[linuxguru]

I eventually went with a Howland Current Pump as the power stage, nested within a traditional voltage-series GNFB loop, i.e. a composite amp with elements borrowed from Mauro's MyRef and Walt Jung's Composite Buffer. I did this on the hunch that it would be easier to stabilize with a transconductance amp as the power stage, and that turned out to be correct for the LM1875 (the key idea to stabilize it is a single small-valued cap or RC between the junction of Rz-R2, and the non-inverting input of the current-pump - which turns out to be an improved variant of the DeBoo integrator). The LM3886 in a nearly identical circuit turned out to be harder to stabilize, but I succeeded after borrowing ideas from the MyRef Rev C.

I looked at the TDA729x, but it had several issues:

1) It is reported to be finicky about power sequencing - it's been known to latch-up on occasion.
2) It requires a bootstrap cap, and I wasn't sure of the effect of the bootstrap (or lack of it) in a Howland Current Pump.
3) It doesn't have an official Spice model to help model stuff like 1) and 2) above.

So I stayed with the known LM1875 and LM3886, with no regrets.

 

[sfthurber]

I looked at the TDA729x, but it had several issues:

1) It is reported to be finicky about power sequencing - it's been known to latch-up on occasion.
2) It requires a bootstrap cap, and I wasn't sure of the effect of the bootstrap (or lack of it) in a Howland Current Pump.
3) It doesn't have an official Spice model to help model stuff like 1) and 2) above.

So I stayed with the known LM1875 and LM3886, with no regrets.

Thanks for posting the heads-up about TDA729x, because I was just starting work on a TDA7293 version. Food for thought about perhaps stick with what I know will work.

 

[linuxguru]

Thanks for posting the heads-up about TDA729x, because I was just starting work on a TDA7293 version. Food for thought about perhaps stick with what I know will work.

The TDA729x family may well work fine, it's just that I had some dead knowledge zones with it. It's entirely possible that those who're familiar with it will find it easy to use in a nested topology.

 

[danielwritesbac]

The TDA7293 will probably behave very nicely if when using TDA7293 specific bootstrap hookup. At higher voltage, it turns out to be a pleasant tone rock the house classic mid-fi, with actually very good clarity and durability. They're also easily paralleled with only one voltage amp active, which is a great simplicity and intensely powerful.

The TDA7294 may behave if used at voltages typical for TDA7296/LM1875. Conveniently, in undervolted condition, TDA7294 supports slightly lower gain, at much benefit to imaging. Scaled even smaller than the LM1875, the TDA7294 can make a very high fidelity low wattage amplifier.

The majority of the trouble with TDA7293/4/5/6 ends if the power decoupling caps are 220u although parallel 220u (440u) may be used for more punchy bass. It is somewhat helpful if the bootstrap cap is 33u rather than 22u. It is rather extremely helpful if the nfb-shunt-cap (in- coupler) is sized up much larger, such as 220u with 2k7 for that RC. Also, the optimal gain factor and feedback current do both depend on the power voltage chosen. For TDA7294, the input cap (in+ coupler) should be 1uF or smaller. For TDA7293, the input cap (in+ coupler) can be whatever size you like (if when using TDA7293 specific bootstrap hookup).

Generally, the standby circuit is hooked up per the datasheet.
However, the mute circuit can be powered through only one part--a simple 10k resistor will do it reliably (no extra resistor, no extra diode, no extra cap needed).

I think that the easiest way to use TDA7293 as a nested amplifier is to use it in slave mode, driven by whatever you like for voltage amp, such as a natsemi driver chip.