In another thread, Armand made me aware of a High Speed, High Voltage 1 Amp Output Drive current, the ADA4870.
By paralleling and bridging it could be a nice device to build a main amp.
Analog Devices has a Spice model, readily available for this CFA.
However when trying to use it, I had great difficulties in getting it to do anythingat all, producing a steady output voltage of either at -16V or +16V depending on the voltage at the Shutdown pin.
To enable the device, this shutdown pin should be at between 1.1 and 5 Volt above Vee. The ON pin is not available in the Spice model, but is internally kept at Vee.
In my case Vee was at -20Volt and V Shutdown at -15V.
Vcc was set at +20Volt, Rf at 1310 Ohm and Rg at 243 Ohm , all as in the specs from Analog, giving a gain of -5.
After much trial and error I found that with a 100K resistor between the Shutdown pin and 5V above Vee, it finally started to produce an output signal, see image below.
Green is the input signal at 10Khz and blue the output current.
But, maximum output current at any output voltage was a humble +/- 80mA.
Supply current at idle was 9mA instead of 33mA, Slew Rate was 18V/usec insteadof 2500V/usec.
And last but not least, input bias current for both inputs was below 0.3uA instead of ca 10uA.
Using the positive input, giving a gain of +6, did not change anything to the above.
So my question:
Has anybody experience with this device and what is the secret to getting the sim working properly.
Hans
By paralleling and bridging it could be a nice device to build a main amp.
Analog Devices has a Spice model, readily available for this CFA.
However when trying to use it, I had great difficulties in getting it to do anythingat all, producing a steady output voltage of either at -16V or +16V depending on the voltage at the Shutdown pin.
To enable the device, this shutdown pin should be at between 1.1 and 5 Volt above Vee. The ON pin is not available in the Spice model, but is internally kept at Vee.
In my case Vee was at -20Volt and V Shutdown at -15V.
Vcc was set at +20Volt, Rf at 1310 Ohm and Rg at 243 Ohm , all as in the specs from Analog, giving a gain of -5.
After much trial and error I found that with a 100K resistor between the Shutdown pin and 5V above Vee, it finally started to produce an output signal, see image below.
Green is the input signal at 10Khz and blue the output current.
But, maximum output current at any output voltage was a humble +/- 80mA.
Supply current at idle was 9mA instead of 33mA, Slew Rate was 18V/usec insteadof 2500V/usec.
And last but not least, input bias current for both inputs was below 0.3uA instead of ca 10uA.
Using the positive input, giving a gain of +6, did not change anything to the above.
So my question:
Has anybody experience with this device and what is the secret to getting the sim working properly.
Hans
Attachments
WFM.
Sounds like your pin assignment may be messed up. Try the attached symbol, based on AD8038. It expects ADA4870.sub in cmp\sub, otherwise edit as needed.
Sounds like your pin assignment may be messed up. Try the attached symbol, based on AD8038. It expects ADA4870.sub in cmp\sub, otherwise edit as needed.
Attachments
Last edited:
There are some cons about ADA4870. It’s very similar with LT1210.
First of - relatively low current capability. So for appropriate audio power amplifier you need at least 8x paralleled chips with very good heatsink mount. Keep in mind underchip powerpad, so mounting cold be very tricky. Of course, doing this will result in relatively huge high-freq PCB, not only priced chips.
Next you could come to output zero compensation around different chips. It’ll better to be done via inverting OPS devices connection with dedicated 2’nd order integrators at noninverting pin of each of them.
Next you could come to composite opamp with something fast and well-audible in front of ADA4870. Try something like LM6171 or AD8065.
Last you could come is sound coloration. Nor hude feedback, nor PCB layout, nor choosed freq compensation couldn’t provide you with aesthetic sounding with poor and low quality OPS devices main capacitors. Use at least ELNA RFS2 or Nichicon KZ/KA series dedicated to each chip.
First of - relatively low current capability. So for appropriate audio power amplifier you need at least 8x paralleled chips with very good heatsink mount. Keep in mind underchip powerpad, so mounting cold be very tricky. Of course, doing this will result in relatively huge high-freq PCB, not only priced chips.
Next you could come to output zero compensation around different chips. It’ll better to be done via inverting OPS devices connection with dedicated 2’nd order integrators at noninverting pin of each of them.
Next you could come to composite opamp with something fast and well-audible in front of ADA4870. Try something like LM6171 or AD8065.
Last you could come is sound coloration. Nor hude feedback, nor PCB layout, nor choosed freq compensation couldn’t provide you with aesthetic sounding with poor and low quality OPS devices main capacitors. Use at least ELNA RFS2 or Nichicon KZ/KA series dedicated to each chip.
Hans
Might be a model issue. Have you tried the LTspice forum on Yahoo? Helmut Sennewald is the guru there and may be able to give you some insight.
Might be a model issue. Have you tried the LTspice forum on Yahoo? Helmut Sennewald is the guru there and may be able to give you some insight.
The problem is of a different nature. The model works fine in LTSpice XVII,but not in LTSpice IV.
You do not even have to install a .ASY file, because all this is written somewhere deep down inside LTSpice XVII.
But thanks anyhow for the support.
This is the first time I experience that a subckt works in XVII but not in IV.
Hans
Bonsai & BesPav,
Thanks for your support.
As it seems my problem is solved by going to LTSpice XVII.
Hans
Thanks for your support.
As it seems my problem is solved by going to LTSpice XVII.
Hans
Attachments
Yes, subckt statement must be uncommented for creating model, but instead commented again for using in LTSpice XVII.
I have the same idea for a long time, and ordered 2 chips for experiment.
2 Objectives: 1. to see if I can stablize it. 2. to see if it sounds good.
Many ICs spec well but sounds terrible, IMO, experiment is a must.
2 Objectives: 1. to see if I can stablize it. 2. to see if it sounds good.
Many ICs spec well but sounds terrible, IMO, experiment is a must.
Nice headphone Amp! Might be difficult to heat sink in a big Class AB amp and the rail voltage is limited so a bridge configuration is probably called for.
- Status
- Not open for further replies.