[minima] Re: RX gain of Bidirectional Amp?
- From: "Joe Rocci" <joe@xxxxxxxxxx>
- To: <minima@xxxxxxxxxxxxx>
- Date: Wed, 15 Oct 2014 07:41:27 -0400
I've had another detailed look at the Rx-side amp in the BiDi. The problem
begins with the choice of a common-base input stage with a resistive load in
the collector. In order to get a nice 50 ohm input impedance, the common base
amplifier has to be biased around 0.5ma collector current. This very low
standing DC current makes it difficult to get the collector voltage down in a
range where there's optimum voltage-swing headroom and a low enough DC voltage
to be directly coupled to the buffer stages. The result is there's very little
voltage swing headroom, and a large base voltage on the emitter-follower
buffers, resulting in lots of current in those stages. In a rig that doesn't
have AGC, it's very important that the IF amplifiers don't overload on strong
signals, so this should be considered. I don't know why a common-base input
stage was chosen here; the most common reasons are :
1) Predictable input impedance determined by bias current
2) Wide bandwidth
3) Excellent reverse isolation
Perhaps Farhan can shed some light on this design decision.
I think the performance requirements for this stage could be better met by
using the same circuit in the Rx side as is used in the Tx side. The gain would
be the same, the strong signal headroom is (allegedly) better, the match is
very good and the isolation should be plenty adequate. If my previous re-bias
suggestion pans out, the current would be much lower too.
BTW Steve, you're correct that noise figure isn't terribly important at the
lower HF frequencies where atmospheric noise predominates, but there are
limits. If you consider 1.5 db loss in the input LPF, 7dB in the mixer, 1.5 dB
in the crystal filter and 5.5 dB NF in the IF amp, you end up with an overall
noise figure of almost 16 dB. That means that, instead of 0.3uv usable
sensitivity on SSB, you get 2.0uv. That might be adequate on the lower end of
the bands where it's noisy anyway, but when you get to the much quieter 10
meter range, you might wish you could dig out the weaker signals.
Joe
W3JDR
----- Original Message -----
From: Joe Rocci
To: minima@xxxxxxxxxxxxx
Sent: Wednesday, October 15, 2014 6:03 AM
Subject: [minima] Re: RX gain of Bidirectional Amp?
Most of the current on the Rx side is in the output buffer, primarily due to
how the whole circuit is biased. I'll take another look and see if there's an
opportunity for improvement. Last time I looked, it was problematic without
adding parts.
Sent from my tablet
-------- Original message --------
From Steve VK2SJA <vk2sja@xxxxxxxxxxx>
Date: 10/14/2014 11:03 PM (GMT-05:00)
To minima@xxxxxxxxxxxxx
Subject [minima] Re: RX gain of Bidirectional Amp?
Thanks Mac,
This is Joe's own suggestion for the TX side of the bidirectional amp. Joe
and I were talking about the other side/direction, the RX side of the amp.
I think Joe was saying he thought that he may have seen somebody else's
suggested mod to reduce current here? At least that's how I was reading
it.
Great "Construction Guide" BTW.
73, Steve.
> Steve,
>
> I've copied Joe Rocci's original post below.
>
> TL'DR:
> Change the 220 Ohm collector resistor to 1000 Ohms to decrease
> the current consumption. Increase the emitter degeneration
> resistor to 20 Ohms to bump the gain up to 20 dB.
>
> 73,
>
> Mac AE5PH
>
> Original message posted by Joe Rocci:
> Hi Steve
>
> Here are some results from modeling the Tx section in LTSpice:
>
> 1) Gain is about 17db @ 20 mhz
> 2) Input match (return loss) is about 16db and pretty flat up to 50 Mhz
> 3) Output match is quite good, at the expense of about 6db gain loss,
> due to the series output matching resistor
> 4) Current draw is about 55ma
>
> Some comments:
> 1) The voltage-amp stage is biased at almost 11 volts on the collector,
> leaving very little headroom for large voltage swings.
> 2) The emitter follower buffer stages draw the lion's share of the DC
> current, but are non-optimum for driving large voltage swings into
> non-resistive loads at high frequencies; that's why so much static DC
> current is required.
> 3) If you increase the 220 ohm collector resistor to 1000 ohms, the
> collector drops closer to 5.5 V, and the overall current consumption
> drops to about 28 ma. If you then increase the emitter degeneration
> resistor to 20 ohms, you end up with about 20 dB gain (@20 Mhz), 22 dB
> input match, and presumably much better output voltage swing headroom. I
> discovered all this while trying to goose the amp up for a little better
> performance at 30 Mhz, anticipating a new crystal filter.
>
> Of course, all of this is 'on paper' and needs to be verified.
>
> Joe
> W3JDR
>
>
> On 10/14/2014 08:22 PM, Steve VK2SJA wrote:
>> Thanks Joe,
>>
>> So ~16 to 20dB in both directions.
>>
>> I had a quick look around the archive this morning. If there is a mod
>> for
>> reducing current in the RX side of the bidirectional amp I've managed to
>> miss it.
>>
>> Anyone know of such a mod?
>>
>> The original W7ZOI bidirectional amp article quotes a noise figure of
>> 5.8dB when the amp was setup for about 16dB of gain. So 4.5dB seems an
>> improvement on that and around the same ball park. So the LTSpice
>> simulation is probably on the money.
>>
>> Last night I was reading EMRFD and learnt how to go about measuring
>> noise
>> properly. I've ordered a LNA and when it arrives I'll try measure it in
>> the real-world and report back.
>>
>> I've been reading that at HF frequencies there is so much atmospheric
>> noise that a 3kHz wide SSB rig will typically see about 14dB of noise.
>> And
>> that as such any amp with a noise factor (NF) of of a bit less than
>> ~14dB
>> is probably adequate for this application. Because you won't see the
>> locally generated amp noise over the atmospheric noise anyway. So while
>> we
>> should always design and build as best we can. From a practical stand
>> point is there anything to be gained by reducing the NF to a figure much
>> below ~6dB?
>>
>> 73, Steve.
>>
>>
>>
>>> Steve
>>>
>>> LTSpice simulation indicates about 16db gain at 20 mhz with the
>>> component
>>> values shown on the web page. The standing current is also very high,
>>> at
>>> over 40ma. I think there was a mod to bring the current down? Also, the
>>> simulated noise figure is about 4.5db, pretty high. However, I've never
>>> checked to see how accurate LTSpice is in this regard.
>>>
>>> Joe
>>>
>>> ----- Original Message -----
>>> From: "Steve VK2SJA" <vk2sja@xxxxxxxxxxx>
>>> To: <minima@xxxxxxxxxxxxx>
>>> Sent: Tuesday, October 14, 2014 6:40 AM
>>> Subject: [minima] RX gain of Bidirectional Amp?
>>>
>>>
>>>> Hi Guys,
>>>>
>>>> The gain of the Minima Bidirectional Amp in the TX direction is ~16 to
>>>> 20dB.
>>>>
>>>> Can someone tell me what the gain is supposed to be for the other half
>>>> of
>>>> the duel Amp, in the RX direction? Same?
>>>>
>>>> 73, Steve.
>>>>
>>>>
>>>>
>>>>
>>>
>>>
>>
>>
>>
>
>
Other related posts:
