AC current source with Howland circuit

Hi Fazlul,

I don't specifically cover the LMH6645, but can make some general comments regarding your application:

• The current accuracy of the Improved Howland Current Pump is directly affected by the exactness of component values and amplifier errors. Be sure to review the following application note regarding the Improved Howland Current Pump: http://www.ti.com/lit/an/snoa474a/snoa474a.pdf
• The LMH6645 is not rated for +/-12 V supplies. The maximum supply-to-supply voltage is 12.6 V. The simulation model may not operate as you would like with +/-12 V supplies.
• 20 mA through a 1 k load is 20 V. Whatever op amp you select it must be able to support that output level at the supply levels you intend to operate it.

Regards, Thomas

PA - Linear Applications Engineering

Hi,

Thank you Thomas for the advice. I studied the advanced Howland circuit current pump to obtained 20mA between frequency range of 1k-500Khz with load resistance between 100-1kohm. With 1kohm load the maximum output voltage is 20V.  I decided to use resistance 100kohm with 0.01% tolerance as the output current and impedance are sensitive to resistors variations. I also have shortlisted some of the op-amp that can be used for the purpose.

LHV870

LME49870

OPA604

I attached simulation results (ac analysis) for the OPA604. The result does not meet my expectation as the output voltage drops before it reaches 100kohm. Can anyone gives some advice on this? How to simulate output impedance for the howland circuit?

 I need to perform simulation for the remaining devices but unfortunately I've not found pspice model for LME49870 and for the LHV870 the device it is not available in my.element14.com. Can anyone provide the PSpice model of LME49870 device? Please help.

Regards

Fazlul

 0841.Howland Circuit ver.1.pdf

Hello Fazlul,

The requirements of +/-24 V supplies and relatively wide bandwidth makes this a tough application for an operational amplifier. Most of the otherwise useable candidates top out with +/-18 V maximum supplies.The only device of the 3 that is from our Precision Amplifier product group is the OPA604. It is an excellent, legacy audio operational amplifier, but I suspect that it is running out of output drive when you attempt to swing the output within 4 V of supply rail while delivering 20 mA, with the +/-24 V supplies. All of its specifications are written around +/-15 V supplies so the performances can be different with the higher supplies. The OPA604 is rated with an Absolute Maximum Power Supply Voltage of +/-25 V. You may find that if your increase the supplies to +/-24.5 V that you get the swing you need. I don't think I would push it any higher.

The LHV870 and LME49870 are handled by TI's SVA amplifiers group (formerly National Semiconductor). They would be best able to assess their amplifier's performance in your application. However, when reviewing the data sheets for these two amplifiers I find that their maximum rated supply voltage is +/-23V, two volts lower on each supply than the OPA604. And, the swing from the supply rails looks to be about 3 to 4 V which will likely be the same limitation you are having with the OPA604.

An operational amplifier simulation model really must have the open-loop output impedance (Zo) designed in to get accurate results. Over about the past 5 to 7 years we have been characterizing and striving to include Zo in our Precision Amplifier operational amplifier simulation models. It is required for a realistic, accurate stability analysis. The OPA604 is a legacy product and at the time it was developed (~1992) the inclusion of Zo was not a consideration. If you take an such as the OPA211, or OPA1641, they will include accurate Zo in their models. Unfortunately, they cannot the handle the +/-24 V supplies you plan to use.

Regards, Thomas

PA - Linear Applications Engineering

Hi all,

Thank you for the advice. I will stick with OPA604 for the Howland circuit. I changed the Vcc and Vee supply and I need assistance to calculate the output impedance of Howland current pump circuit as shown in Fig.1. I  performed simulation on the output impedance as in Fig.2 but I'm not sure if the result is correct. The target is to have output current 20mA with high output impedance between 1k-500kHz The output impedance calculated according to the formula:

Z=[(R6a+R7)*I1 - (R6b+R7)*I2]/ (I1-I2)

 The R6a and R6b are two resistors value at 100ohm and 500ohm recpectively.R7 is fixed value at 500ohm.The output impedance was calculated from the difference in the current flowing using R6a and R6b. Can anyone tell me whether the result is correct or not? If not please guide me to find the correct result. Thank you in advance.

Regards

Fazlul

6138.Howland Circuit ver.2.pdf

See attached PowerPoint for analysis of Zout of your Improved Howland Current Pump.  In addition the attached two articles in the following posts may be of interest regarding output impedance and the Improved Howland Current Pump.

Improved Howland Zout Article 1.

Improved Howland Article 2

Follow-on Analysis - see Rev A attached and be sure to note dip in Zout of Howland over frequency.

Hi Tim,

Thank you for the valuable resources. I really appreciate it. Now I have designed the Howland circuit and simulated the output impedance of the circuit. So far I found that LM7171 gives the best results, however I still not satisfied with the results especially at 500kHz. I would expected the Zo will above 100k. The current results shows at 500kHz the Zo is 45kohm. I've tried several high values with high tolerances but the result still less than 50kohm. Any advice how to improve the results? 

Regards

Fazlul

6406.Howland Circuit ver.3.pdf